Title

Game-theoretic coordination and configuration of multi-level

supply chains

Advisor(s)

Huang, GQ

Author(s)

Huang, Yun;

Citation

Issued Date

URL

Rights

2010

http://hdl.handle.net/10722/130938

The author retains all proprietary rights, (such as patent rights)

and the right to use in future works.

GAME-THEORETIC COORDINATION AND

CONFIGURATION OF MULTI-LEVEL SUPPLY CHAINS

By

Yun Huang

B.Sc. SEU
M.Sc. ZJU

A thesis submitted in partial fulfilment of the requirements for

the Degree of Doctor of Philosophy
at The University of Hong Kong

June 2010

Abstract of thesis entitled

Game-theoretic Coordination and Configuration of

Multi-level Supply Chains
Submitted by
Yun Huang
for the degree of Doctor of Philosophy
at The University of Hong Kong
in June 2010
This research is concerned with coordinating supply chain configuration decisions,
i.e., pricing, inventory and supplier selection, in multi-level supply chains. Game
theory is the main coordination approach employed to model and study these
problems. This research has made significant contributions using several scenarios of
supply chain configuration.
The first scenario coordinates pricing decisions in a three-level supply chain
composed of a single supplier, a single manufacturer and a single retailer. We consider
the effects of different power structures, channel structures and market related
parameters on the optimal prices and profitability of the chain members. The results
show that the manufacturer had better take the channel leadership. The integration for
the manufacturer and the retailer cannot always improve their profits even in a
monopoly. When product cost is larger than a certain level, all the chain members
could benefit from the increase of price elasticity.
The second scenario investigates the pricing and inventory coordination problem
in a two-level supply chain with one single manufacturer and multiple competing

retailers. The problem is modeled as dynamic non-cooperative games. Analytical and

computational methods are employed to solve this game model. The third scenario
further studies pricing and inventory coordination problem in a multi-level supply
chain and extends the above dynamic non-cooperative game into a three-level nested
Nash game. Through numerical experiment, we show that when one retailers market
becomes more sensitive to the price, his profit will decrease but the other retailers will
benefit. Suppliers profits may increase as the manufacturers setup cost increases.
When one suppliers raw material cost increases, the other suppliers profits might
decrease more significant.
The fourth scenario further incorporates supplier selection into pricing and
inventory coordination problem. The manufacturer has also to choose the optimal
components of certain functionality from his alternative suppliers to produce the
products for the retailers. A numerical study in computer industry shows that when
one retailers market scale increases, the manufacturer tends to use high-end
components for the product sold to this market. As a result, the manufacturer would
benefit from this strategy and the supply chain system includes fewer number of
suppliers and their total profit becomes lower. The manufacturer would setup his
production more frequently when the component selection strategy is employed.
The last scenario proposes an integrated approach for optimal supplier selection,
pricing and inventory decisions in a multi-level supply chain. We formulate a
mixed-integer programming model to optimize the decisions and use cooperative

game approach to evaluate the supply chain members. The numerical results show
that in an integrated manner, the increase of one market scale will benefit the product
in this market and not influence the other retail market. The increased market scale
also enhances the importance of this market and reduces the importance of the
suppliers provided components not used for the product in this market.

(460 words)

DECLARATION

I declare that this thesis represents my own work, except where due
acknowledgement is made, and that it has not been previously included in a thesis,
dissertation or report submitted to the University of Hong Kong or to any other
institution for a degree, diploma or other qualifications.

Signed
Huang, Yun

ACKNOWLEDGEMENTS

My first and deepest gratitude must go to my supervisor, Prof. George Q. Huang

for his outstanding supervision, wisdom advice, and kind assistance during my
doctoral study. His seriousness, enthusiasm and immense knowledge towards
scientific research not only help my research but also edify my attitude towards work
and life. Actually, no words can adequately express my appreciation to him.
I would also like to thank my dearest colleagues in our research group for their
suggestions and helpful inspirations in my research work. All the lab mates at Room
823 and 824 made here a convivial place to work. I wish to thank all of them for their
friendship and support in my daily life over the years.
My wordless thanks go to my families for their understanding, encouragement and
enduring support through my life. This dissertation is simply impossible without
them.

ii

CONTENTS

Declaration ..................................................................................................................... i
Acknowledgements .........................................................................................................ii
Contents ....................................................................................................................... iii
List of Figures .............................................................................................................. vii
List of Tables ............................................................................................................... viii
List of symbols............................................................................................................... ix
CHAPTER 1 Introduction .............................................................................................. 1
1.1 Background and motivation .............................................................................. 1
1.2 Research scope .................................................................................................. 3
1.3 Scenario design and research objective ............................................................ 7
1.3.1 Motivating scenarios ............................................................................... 7
1.3.2 Research objectives ................................................................................. 9
1.4 Challenge and significance ............................................................................. 10
1.4.1 Scenario I: CCSC with pricing decision ............................................... 10
1.4.2 Scenario II: CCSC with competing retailers ......................................... 12
1.4.3 Scenario III: CCSC with price and inventory coordination .................. 12
1.4.4 Scenario IV: CCSC with supplier selection .......................................... 13
1.4.5 Scenario V: CCSC with integration ...................................................... 15
1.5 Research framework ....................................................................................... 15
1.6 Thesis organization ......................................................................................... 19
CHAPTER 2 Literature review .................................................................................... 22
2.1 Introduction ..................................................................................................... 22
2.2 Supply chain coordination .............................................................................. 23
2.2.1 Coordination of pricing and inventory decisions .................................. 24
2.2.2 Coordination on other supply chain configuration decisions ............... 33
2.2.3 Coordination of multi-level supply chains ............................................ 36
2.3 Game theory in scm problems ........................................................................ 37
2.3.1 Pricing game model .............................................................................. 38
2.3.2 Cooperative game ................................................................................. 43
2.4 Supply chain configuration ............................................................................. 45
iii

2.5 Summary ......................................................................................................... 48

CHAPTER 3 Game-theoretic Pricing Coordination in a Multi-level Supply Chain with
Different Channel Structures ....................................................................................... 51
3.1 Introduction ..................................................................................................... 51
3.2 Model formulation and notations .................................................................... 54
3.3 Decentralized channel ..................................................................................... 56
3.3.1 Manufacturer Stackelberg ..................................................................... 57
3.3.2 Vertical Nash ......................................................................................... 59
3.4 Semi-integrated channel.................................................................................. 60
3.4.1 MR-integration channel ........................................................................ 60
3.4.2 SM-integration channel ......................................................................... 63
3.5 Integrated channel ........................................................................................... 65
3.6 Discussion ....................................................................................................... 66
3.6.1 Effects of power structure ..................................................................... 66
3.6.2 Effects of channel structure .................................................................. 68
3.6.3 Effects of market related parameters .................................................... 72
3.7 Summary ......................................................................................................... 78
CHAPTER 4 Game-theoretic Coordination of Pricing and Inventory Decisions in a
Supply Chain with Retailer Competition ..................................................................... 84
4.1 Introduction ..................................................................................................... 84
4.2 Assumptions and notations ............................................................................. 87
4.2.1 Problem description and assumptions ................................................... 87
4.2.2 Notations ............................................................................................... 88
4.3 Model formulation .......................................................................................... 90
4.3.1 Game scheme ........................................................................................ 90
4.3.2 The retailers model .............................................................................. 92
4.3.3 The manufacturers model .................................................................... 94
4.3.4 MR game model .................................................................................... 97
4.4 Solution algorithm .......................................................................................... 97
4.4.1 Reaction functions ................................................................................ 98
4.4.2 Algorithm for the equilibrium ............................................................. 104
4.5 Numerical example and sensitivity analysis ................................................. 105
4.6 Summary ....................................................................................................... 107

iv

CHAPTER 5 Game-Theoretic Coordination of pricing and Inventory Decisions in a

Multi-level Supply Chain ........................................................................................... 111
5.1 Introduction ................................................................................................... 111
5.2 Problem statement and notations .................................................................. 114
5.2.1 Problem description and assumptions ................................................. 114
5.2.2 Notations ............................................................................................. 116
5.3 Model formulation ........................................................................................ 118
5.3.1 Game scheme ...................................................................................... 118
5.3.2 The retailers model ............................................................................ 121
5.3.3 The manufacturers model .................................................................. 123
5.3.4 The suppliers model ........................................................................... 125
5.3.5 Three-level nested Nash game model ................................................. 126
5.4 Solution algorithm ........................................................................................ 126
5.4.1 Reaction functions .............................................................................. 127
5.4.2 Algorithm for the three-level nested Nash equilibrium ...................... 131
5.5 Numerical example and sensitivity analysis ................................................. 133
5.5.1 Simulation results................................................................................ 134
5.5.2 Managerial implications...................................................................... 136
5.6 Summary ....................................................................................................... 138
CHAPTER 6 Game-Theoretic Coordination of Pricing and inventory Decisions
considering Supplier Selection in a Multi-level Supply Chain .................................. 140
6.1 Introduction ................................................................................................... 140
6.2 Problem description and notations ................................................................ 143
6.3 The mathematical model ............................................................................... 147
6.3.1 Game scheme ...................................................................................... 147
6.3.2 The mathematical model ..................................................................... 148
6.4 Solution algorithm ........................................................................................ 154
6.4.1 Reaction functions .............................................................................. 154
6.4.2 Algorithm for the Nash equilibrium of RMS game ............................ 159
6.5 Numerical results .......................................................................................... 160
6.5.1 Initialization ........................................................................................ 160
6.5.2 Simulation Results .............................................................................. 162
6.5.3 Managerial implications...................................................................... 167

6.6 Summary ....................................................................................................... 169

CHAPTER 7 Integrated supplier selection, pricing and inventory decisions in a
multi-level supply chain ............................................................................................. 176
7.1 Introduction ................................................................................................... 176
7.2 Problem description ...................................................................................... 179
7.3 Mathematical models formulation ................................................................ 181
7.3.1 Notations ............................................................................................. 181
7.3.2 Integrated supply chain model ............................................................ 183
7.3.3 Cooperative game model formulation ................................................ 185
7.3.4 Solution procedure .............................................................................. 187
7.4 Numerical results .......................................................................................... 187
7.4.1 Results for base example .................................................................... 188
7.4.2 Sensitivity analysis.............................................................................. 189
7.4.3 Managerial implications...................................................................... 192
7.5 Summary ....................................................................................................... 193
CHAPTER 8 Conclusion and future research ............................................................ 198
8.1 Major achievements and contributions ......................................................... 198
8.1.1 CCSC with pricing decision................................................................ 198
8.1.2 CCSC with competing retailers .......................................................... 200
8.1.3 CCSC with price and inventory coordination ..................................... 201
8.1.4 CCSC with supplier selection ............................................................. 202
8.1.5 CCSC with integration ........................................................................ 204
8.2 Limitations and directions for future work ................................................... 205
8.2.1 Limitation and immediate extensions for scenarios in this thesis....... 205
8.2.2 Game models ...................................................................................... 209
8.2.3 Supply chain configuration ................................................................. 211
8.3 Overall conclusion ........................................................................................ 212
Appendix .................................................................................................................... 215
Reference ................................................................................................................... 217
Publications during Ph.D. study................................................................................. 231

vi

LIST OF FIGURES

Figure 1.1. Motivating scenarios of CCSC problem...................................................... 7

Figure 1.2. Research framework .................................................................................. 16
Figure 3.1. Game structures for the decentralized channel .......................................... 57
Figure 3.2. Game structures for semi-integrated channels .......................................... 62
Figure 4.1. MR game structure .................................................................................... 91
Figure 4.2. Inventory levels of retailer and manufacturer ........................................... 96
Figure 5.1. Three-level nested Nash game structure.................................................. 120
Figure 5.2. Inventory levels of retailer, manufacturer and supplier........................... 122
Figure 6.1. Coorination of supplier selection, pricing and inventory decisions (CSSPI):
an illustration ........................................................................................... 144
Figure 6.2. Inventory levels of retailer and manufacturer ......................................... 149
Figure 6.3. Influence of market scale parameter on product wholesale prices .......... 162
Figure 6.4. Influence of market scale parameter on profits for chain members ........ 164
Figure 6.5. Influence of component cost on product wholesale prices ...................... 166
Figure 6.6. Influence of component cost on profits for chain members .................... 167
Figure 7.1. Manufacturers inventory level for final product .................................... 185
Figure 7.2. Shapley value for different market scale parameters .............................. 191
Figure 7.3. Shapley value for different component costs .......................................... 191

vii

LIST OF TABLES

Table 2.1. References for pricing and inventory coordination .................................... 29

Table 2.2. Game models for supply chain coordination .............................................. 39
Table 3.1. Results for leader-follower power structure ............................................... 80
Table 3.2. Results for independent power structure .................................................... 81
Table 3.3. Effects of market scale parameter a ............................................................ 82
Table 3.4. Effects of price elasticity b ......................................................................... 83
Table 4.1. Results for the manufacturer and the retailers .......................................... 109
Table 5.1. Results for suppliers, manufacturer and retailers...................................... 139
Table 6.1. Parameters and variables of retailers ........................................................ 148
Table 6.2. Parameters and variables of manufacturer ................................................ 150
Table 6.3. Parameters and variables of suppliers ....................................................... 152
Table 6.4. Results for suppliers under different market scale.................................... 171
Table 6.5. Results for manufacturer and retailers under different market scale ........ 172
Table 6.6. Results for suppliers under different production cost ............................... 173
Table 6.7. Results for manufacturer and retailers under different production cost ... 173
Table 6.8. Results for suppliers under different component cost .............................. 174
Table 6.9. Results for manufacturer and retailers under different component cost ... 175
Table 7.1. Profits for all valid coalitions.................................................................... 196
Table 7.2. Decisions and Shapley values ................................................................... 197
Table 8.1. Summary of CCSC problem in this thesis ................................................ 200

viii

LIST OF SYMBOLS

A constant in the demand function of the retailer, which represents the

market scale
A constant in the demand function of retailer l, which represents his market

Arl

scale
b

Coefficient of the products price elasticity for the retailer

cLij

The cost of component Lij paid by supplier

cm

Production cost per unit product

cml

Production cost per unit product l

cslv

The purchasing cost of raw material vl used for product l

csv

Raw material cost paid by supplier v

Retailers annual demand

Drl

Retailer ls annual demand

erll

Coefficient of the products price elasticity for retailer l

erlj

Coefficient related to the degree that product j could be substitutable by

product l

FLij

Fixed cost of using component Lij

Gml

Manufacturers profit margin for product l

Grl

Retailer ls profit margin

Gsv

Supplier vs profit margin

hmpl

Holding costs per unit of product l

hmrs

Holding costs per unit of raw material purchased from supplier v

hmslv

Holding costs per unit of raw material vl used for product l

hrl

Holding costs per unit of product l for retailer l

hsv

Holding costs per unit of raw material inventory for supplier v

krl

The integer divisor used to determine the replenishment cycle of retailer l

K sv

The integer multiplier used to determine the replenishment cycle of supplier

v
L

Total number of retailers

Index of the manufacturer

mm

Manufacturers profit margin

ix

List of symbols

mr

Retailers profit margin

nml

The integer multiplier used to determine the procurement cycle of raw

materials used for product l

Oml

Ordering processing cost of raw materials used for product l per order

Orl

Ordering processing cost for retailer l per order of product l

Osv

Order processing cost for supplier v per order

Pl

Annual production capacity product l, which is a known constant

pLij

Decision variable, the price of component Lij charged by the supplier to the
manufacturer

pm

Wholesale price charged by the manufacturer to the retailer

pml

Wholesale price charged by the manufacturer to the retailer l

pr

Retail price charged to the customers by the retailer

prl

Retail price charged to the customers by retailer l

ps

Raw material price charged by the supplier to the manufacturer

psv

Raw material price charged by the supplier to the manufacturer

Qsv

Set of components supplied by supplier v

Index of the retailer

rl

Index of retailer l

Rm

Manufacturers annual fixed costs for the facilities and organization for the
production of this product

Rrl

Retailer ls annual fixed costs for the facilities and organization to carry this
product

Rsv

Supplier vs annual fixed costs for the facilities and organization to carry the
components

Index of the supplier

slvl

Index of raw material vl used for product l, vl =1,2,,Vl

Sm

Setup cost per production

sv

Index of supplier v, v=1,2,,V

Manufacturers setup time interval

tsv Lij

Binary decision variable to indicate whether component Lij is supplied by

supplier v

uLik l

Usage amount of unit component Lij per unit product l

Total number of suppliers

List of symbols

Vl

Total number of raw materials used for product l

wsv

The fixed cost of using supplier v, covering supplier certification, contract

setup, etc.

Xm

Strategy set of the manufacturer. xm X m is his strategy

X rl

Strategy set of retailer l. xrl X rl is his strategy

X sv

Strategy set of supplier v. xsv X sv is his strategy.

z Lij

Binary decision variable to indicate whether component Lij is used

Zm

Payoff function of the manufacturer

Z rl

Payoff function of retailer l

Z sv

Payoff function of supplier v

Usage amount of unit raw material per unit product

Usage of raw material vl for a unit product l

Usage of supplier vs raw material to produce a unit product l

sv

Number of different types of components supplier v is capable of supplying

ijk

Binary decision variable to indicate whether component Lij has been used

lvl

to replace Lik

Binary decision variable to indicate whether supplier v is used

Binary decision variable to indicate whether retailer l is used

Manufacturers profit function

Retailers profit function

Suppliers profit function

xi

CHAPTER 1
INTRODUCTION

1.1 BACKGROUND AND MOTIVATION

The term supply chain management appears first in 1982 (Oliver and Michael
1992). For the last two decades, supply chain management (SCM) becomes an area
that received great attention in business community. One of the most popular
definition is developed by Simchi-Levi et al. (2000):
Supply chain management is a set of approaches utilized to efficiently coordinate
and integrate suppliers, manufacturers, warehouses, and stores, so that merchandise
is produced and distributed at the right qualities, to the right location, and at the right
time, in order to minimize system-wide costs while satisfying service level
requirements.
According to this definition, we can identify three integral parts of the overall
supply chain management process supply chain network, supply chain configuration,
and supply chain coordination.
A supply chain can be regarded as a network of geographically distributed and
administratively decentralized business partners (e.g., suppliers, manufacturer, etc). In
such a decentralized supply chain, decisions of individual partners are often not
coordinated with each other. Their local objectives are often inconsistent with those of
the entire system objectives. As a result, the supply chain becomes less competitive

Chapter 1

(Porter 1985). SCM is an efficient new approach to managing the business processes
from original suppliers to end users.
Supply chain configuration is one of the principal SCM decisions (Chandra and
Grabis 2007). According to Chandra and Grabis (2007), supply chain configuration
decisions can be classified into two types, i.e., macro and micro. At the macro level, it
concerns the whole supply chain system and the decisions of all entities involved. At
the micro level, issues concerning implementation of the macro level decisions are
addressed in supply chain configuration. For example, at the macro level, supply
chain configuration may concern how much to invest in new or existing plants and at
what location. At the micro level, example problems include how much plant capacity
should be allocated to a particular product, etc. We can see that supply chain
configuration covers a large scale of decisions in supply chain management.
The above definition also stresses coordination to be a core requirement of SCM.
The coordination of the supply chain configuration decisions means an approach that
optimizes the system rather than individual elements, improving both the efficiency of
the supply chain and of the firm (Romano 2003; Chan et al. 2004). Recent years have
witnessed the successful implement of supply chain coordination in many large, high
performing focal firms (e.g., Dell, Amazon.com, Wal-Mart, Compaq). For example,
Dell Company outsourced some of its production to a number of OEMs (i.e., Original
Equipment Manufacturers) to convert a low-margin direct sales operation into high
profit, high service business and thus, increased the coordination within its supply

Chapter 1

chain (Mc Williams 1997).

However, the coordination of supply chain configuration decisions is still in the
early stages for SCM (Chan et al. 2004). This is primarily due to the hybrid
environment embraced where individual firms face various supply chain
configuration decisions simultaneously (e.g., pricing, supplier selection, etc.) as well
as the complicated interactions with the other entities of a supply chain network. Due
to the complexity, many researchers (e.g., Choi 1996; Abad 1997; Boyaci and Gallego
2002) focus the coordination problem on two-echelon supply chains, i.e., supplier (or
seller) and retailer (buyer). In reality, a supply chain generally consists of multiple
firms with different functions, such as suppliers, manufacturers, retailers, etc. Such a
situation motivates this research to consider the coordination and configuration of
multi-level supply chain (CCSC) problem.

1.2 RESEARCH SCOPE

From the above discussion, we can see that supply chain configuration covers a
large scale of decision-making problems in supply chain management. Problems that
are commonly discussed in supply chain configuration are as follows (Chandra and
Grabis 2007):
Network design. It is a common way to represent a supply chain as a network (Wu and
OGrady 2001). Thus, it is the core problem of the supply chain configuration
problem.

Chapter 1

Sales and distribution. Ordering and marketing activities are also important supply
chain configuration problems. For example, if a configuration model is used to make
decisions when opening distribution facilities in new markets, then the cost for
attracting customers should be considered.
Logistics design. This problem is referred to deciding on the inventory and
distribution approach, such as multiple integer replenishment mechanism,
vendor-managed inventory, third party logistics, transportation mode, capacity, etc.
Purchasing. In the context of configuration problem solving, purchasing decision is
concerning which suppliers to use for specific materials and in what quantities are
addressed. Other purchasing conditions, such as delivery price and lead time are
determined.
Corresponding to the commonly discussed supply chain configuration problems,
the typical supply chain configuration decisions include, but limited to pricing,
advertising, inventory, service time, ordering policy, supplier selection, etc.
This thesis will focus on three aspects - pricing, inventory and supplier selection
decisions. Pricing is an important marketing strategy related to product positioning
and affects other marketing mix elements, such as channel decisions, promotion, etc
(Michael 2002).
Initially, many firms and researchers focus on one supply chain configuration
decision pricing alone as a tool to improve profits. However for manufacturing
industries, the coordination of price decisions with other aspects of supply chain

Chapter 1

configuration decisions such as production and distribution is not only useful, but is
essential (Weng 1995; Chan et al. 2004). This integration and coordination of pricing,
inventory decisions in retail or manufacturing environments has the potential to
radically improve supply chain efficiencies (Chan et al. 2004).
The manufacturing firms have long acknowledged the importance of keeping
close ties with their downstream customers or retailers. Their attention has extended
upstream as well. Effectively selecting and evaluating the suppliers and managing
their involvement in critical supply chain activities enable manufacturers to achieve
the four dimensions of customer satisfaction: competitive pricing, product quality,
product variety and delivery service (Morgan 1996; Poirier 1999; Michael and Chong
2001).
In this condition, the major issue of this research focuses on coordination of
supply chain configuration decisions in multi-level supply chains. The specific
context of the supply chain network is composed of three echelons supplier tier,
manufacturer tier and retailer tier. Separate studies have been extensively conducted
on coordination of manufacturer-supplier relationship or manufacturer-retailer
relationship (e.g., Huang et al. 2007, Yu et al 2009a, b). However, the linkage of the
two parts not only connects the purchasing decisions (such as supplier selection) and
marketing strategies for customers (such as end-product pricing), but also provides an
approach for optimization of the supply chain from a system perspective.
To efficiently model and analyze decision making in such multi-person situation,

Chapter 1

often with conflicting objectives, game theory is a naturally choice. Osborne and
Rubinstein (1994) define game theory as a bag of analytical tools designed to help
us understand the phenomena that we observe when decision-makers interact. As
such, game theory deals with interactive optimization problems. The essential feature
of these problems is the decision makers are rational to pursue well-defined payoffs,
while taking the other decision-makers behaviors or rational reactions into account.
Cachon and Netessine (2004) comprehensive review the application of game
theory to study supply chain coordination problems. These problems usually focus on
coordination of two-stage supply chain, such as supplier-retailer relationship (e.g.,
Weng 1995), or manufacturer-retailer relationship (e.g., Boyaci and Gallego 2002).
Obviously, a supply chain is not ended up with two echelons and its scope is much
larger. In this research, we will employ game-theoretic approaches to coordinate the
supply chain configuration decisions within the context of multi-level supply chain.
Non-cooperative and cooperative games are two approaches used to coordinate
supply chain. The two approaches are just two different ways to look at the same
problem (Nagarajan and Soi 2006). In the first approach, the game participants
make their decisions to maximize their own objective. In the second one, the goal of
the participants is to maximize their total objective and divide the extra values.

Chapter 1

1.3 SCENARIO DESIGN AND RESEARCH OBJECTIVE

1.3.1 Motivating scenarios

As this thesis focuses on coordination supply chain configuration decisions in

multi-level supply chain, three dimensions that influence the coordination process are
of our concern. They are supply chain network, supply chain configuration decisions

Pricing

Inventory

Supplier
selection

p
up

...

Multiple suppliers, single manufacturer

multiple retailers

Single supplier, single manufacturer

single retailers

Single supplier Multiple retailers

How do we coordinate the supply chain

configuration decisions?
What are the optimal supply chain
configuration decisions?
What are influences of different
parameters on the optimal configuration
decisions and profitability of the supply
chain members?
...

Single supplier single retailer

Cooperative game
Non-cooperative game

Supply chain coordination

and approaches for supply chain coordination, as shown in Figure 1.1.

ch
ly

Advertising

o
tw
ne
n
ai

rk

Supply chain configuration decisions

Figure 1.1. Motivating scenarios of CCSC problem

Many early researchers study supply chain coordination problem in simple

channel which composed of single supplier and single retailer (Jeuland and Shugan
1983). Gradually, competition is introduced between supply chain members. Some
representative work includes Lal and Staelin (1984), Choi (1991; 1996), etc. These
supply chain networks they focus on are generally composed of one/ two suppliers
7

Chapter 1

and one/ two retailers. Many researchers extend such supply chain networks in to one
supplier and multiple retailers case or multiple suppliers and single retailer case
(Huang et al. 2007; Yu et al. 2009). More recently, complex supply chain networks are
also studied. For example, Haq and Kannan (2006) study a supply chain network with
multiple suppliers, multiple distributors, multiple wholesalers and multiple retailers.
The supply chain network dimension is pointed by the inward arrow in Figure 1.1.
Along this arrow, the complexity of supply chain network increases.
The dimension of supply chain configuration decisions is indicated by the
rightward arrow in Figure 1.1. Supply chain configuration decisions are as Section 1.2
introduces, including pricing, inventory, supplier selection, etc. Along the arrow,
increasing supply chain configuration decisions would be involved in this research.
As the upward arrow shown in Figure 1.1, non-cooperative and cooperative games
are two approaches for supply chain coordination.
From Figure 1.1, we can see that there would be various scenarios that combine
the elements of three dimensions could be discussed in supply chain management.
Due to the complexity of the coordination of multi-level supply chain, in this thesis,
we will investigate CCSC problem on coordinating three configuration decisions
pricing, inventory and supplier selection in four types of supply chain networks. Five
scenarios are designed to study the coordination problem. They are: (I) CCSC with
pricing decision; (II) CCSC with competing retailers; (III) CCSC with price and
inventory coordination; (IV) CCSC with supplier selection; (V) CCSC with
8

Chapter 1

integration. Firstly, pricing decision coordination will be analyzed in a relative simple

multi-level supply chain with one single supplier, single manufacturer and single
retailer. Then we will consider the scenario that coordinating pricing and inventory
decisions in two-echelon supply chain with one manufacturer and multiple retailers.
Following these two scenarios, we will extend our attention to pricing and inventory
coordination in multiple suppliers, one manufacturer, and multiple retailers supply
chain. In the fourth and fifth scenarios, the supply chain of multiple alternative
suppliers, one manufacturer and multiple retailers with pricing, inventory, and
supplier selection decisions coordination will be involved.
1.3.2 Research objectives

The primary aim of this research is to effectively and efficiently coordinate

pricing, inventory and supplier selection decisions in multi-level supply chains by
game-theoretic approaches. More specifically, we are concerned with achieving the
following objectives for the overall research as well as the specific scenarios:
(1) Establish an overall research framework and set up research methodology for
coordinating supply chain configuration decisions for multi-level supply chain.
(2) Develop pricing coordination models for multi-level supply chain with one
supplier, one manufacturer and one retailer and study the effects of different
power structures and channel structures on the optimal pricing decisions.
(3) Investigate pricing and inventory coordination game models for two-level supply
chain with one manufacturer and multiple retailers and three-level supply chain
9

Chapter 1

with multiple suppliers, one manufacturer and multiple retailers and study the
influences of different parameters on the optimal pricing and inventory decisions.
(4) Incorporate supplier selection decision into pricing and inventory coordination
problem in multi-level supply chain with multiple alternative supplier, one
manufacturer and multiple retailers, and identify the influences of different
parameters on the optimal pricing, inventory and supplier selection decisions.
The methods investigated in this research will help managers to determining the
optimal supply chain configuration decisions and understand the interaction between
supply chain members as well as the interaction between the market environment and
the optimal configuration decisions.

1.4 CHALLENGE AND SIGNIFICANCE

1.4.1 Scenario I: CCSC with pricing decision

This scenario considers price coordination problem in a three-level supply chain

composed of a single supplier, a single manufacturer and a single retailer. Three types
of channel structures are considered in this work. The first is the decentralized
channel that the manufacturer uses independent supplier and retailer, each of them
optimizing his own profit individually and non-cooperatively. The second channel is
that the manufacturer backward integrates with the supplier or forward integrates with
the retailer and uses independent retailer or supplier. This channel is called by
semi-integrated channel. The third one is the integrated channel structure, where the

10

Chapter 1

manufacturer, the supplier and the retailer operate in total cooperation in a full vertical
integration system. Two power balance scenarios are considered for the decentralized
channel and the semi-integrated channel. The first leader-follower power structure,
where the manufacturer always takes the leadership, is modeled as a Stackelberg
game; the second independent power structure is treated as a Nash game, for the equal
status of the chain members.
The challenge is therefore how to coordinate the price decisions for each supply
chain member under the above power structures and channel structures. On the other
side, to identify the effects of different power structures and channel structures on the
price decisions is another challenge faced.
To consider the above challenges in the context of the three-level supply chain in
this scenario, we use several game models to address the following specific questions:
(1) how will the supply chain members of different echelons make their optimal
pricing decisions under different power structures and channel structures? (2) which
power structure will be preferred by the supply chain members and the entire chain
system under the decentralized and the semi-integrated channels? (3) will the
integration between the manufacturer and the retailer always be profitable? If not,
when would the manufacturer like to choose to forward integrate with the retailer? (4)
how will the market related parameters influence the channel members equilibrium
prices and profits?

11

Chapter 1

1.4.2 Scenario II: CCSC with competing retailers

To coordinate pricing and inventory decisions, in this scenario, we investigate the

pricing and inventory coordination problem in a two-level supply chain with one
single manufacturer and multiple competing retailers. The manufacturer purchases
different raw materials and produces substitutable products for different retailers. The
manufacturer and the retailers are willing to optimize their pricing and inventory
decisions to maximize their individual profits. The problem is modeled as dynamic
non-cooperative game, where the competing retailers formulate a subgame and play
the whole game with the manufacturer.
To formulate the proper game model to simulate the competition between the
retailers and between the manufacturer and retailers as well as determine the optimal
pricing and inventory decisions is the major challenge faced by this chapter. We raise
the following question for retailers: (1) what are the optimal retail prices for products
sold to different markets? (2) how frequent should the product procurement take place?
For the manufacturer, we have to decide on: (1) what are the optimal product
wholesale prices for different retailers? (2) how frequent should the production
process take place? (3) how frequent should the raw material procurement take place?
1.4.3 Scenario III: CCSC with price and inventory coordination

We extend Scenario II into a three-level supply chain coordination problem on

optimal pricing and inventory decisions in a multi-level supply chain, involving
multiple suppliers, a single manufacturer and multiple retailers. The manufacturer
12

Chapter 1

purchases different types of raw materials from his suppliers to produce products for
different independent retailers. The suppliers, the manufacturer and the retailers have
to make decisions on the optimal pricing and inventory decisions respectively to
maximize their profits.
The challenge of this scenario lies in developing the multi-level inventory model
and formulating the game model of the multi-level supply chain to determine pricing
and inventory decisions. We adopt the integer multipliers mechanism as
replenishment mechanism for the raw materials and products and use a three-level
nested Nash game to model this optimization problem. In this game, all the suppliers
formulate the bottom-level Nash game, the whole supplier sector plays the
middle-level Nash game with the manufacturer, and both sectors as a group player
formulate the top-level Nash game with the retailers. Several research questions will
be addressed in this scenario. As to each supplier, we are particularly interested in: (1)
what is the optimal raw material price charged for the manufacturer? (2) how frequent
should the raw material procurement take place? For the manufacturer, he has to
answer: (1) what are the optimal product wholesale prices for different retailers? (2)
how frequent should the production process take place? For the retailers, there are: (1)
what are the optimal retail prices for products sold to different markets? (2) how
frequent should the product procurement take place?
1.4.4 Scenario IV: CCSC with supplier selection

Supplier involvement into product development and critical supply chain activities
13

Chapter 1

brings the manufacturers benefits on the selection of best product variants and
production process (Ragatz et al. 1997, Michael and Chong 2001). This scenario is
not only concerned with the coordination of pricing and inventory decisions, but also
concerned with the suppliers and components selection in a multi-level supply chain
composed of multiple alternative suppliers, a single manufacturer and multiple
retailers. The manufacturer purchases optional components of certain functionality
from his alternative qualified suppliers to produce a set of products to meet the
requirements from the retailers in different markets. The problem is also modeled as a
three-level nested Nash game. The settings of the game scheme is similar to the
previous scenario, where all the competing alternative suppliers formulate the
bottom-level Nash game, the whole supplier sector plays the middle-level Nash game
with the manufacturer, and both sectors as a group player formulate the top-level
Nash game with the retailers.
The main challenge of this scenario is how to integrate supplier and component
selection decision into pricing and inventory coordination problem in multi-level
supply chain. More specifically, for the suppliers, they have to answer: what are their
optimal pricing decisions for the components supplied? The questions faced by the
manufacturer are: (1) what are the optimal suppliers and components that should be
selected? (2) what are the optimal product wholesale prices for different retailers? (3)
how frequent should the production process take place? For the retailers, they are
interested in: (1) what are the optimal retail prices for products sold to different

14

Chapter 1

markets? (2) how frequent should the product procurement take place?
1.4.5 Scenario V: CCSC with integration

This scenario proposes an integrated approach for optimal supplier selection,

pricing and inventory decisions in a multi-level supply chain. The manufacturer
obtains components from the potential qualified suppliers and produces final products
for retailers in different markets. The suppliers, the manufacturer and the retailers
would like to be involved in an integrated manner to determine the pricing, inventory
and supplier selection decisions to maximize the overall supply chain system profit.
We formulate a mixed-integer programming model to optimize supplier selection,
pricing and inventory decisions. A cooperative game approach is used to evaluate the
marginal contribution of the supply chain members. The challenges of this scenario
are the formulation of the cooperative game model and how to use it for supply chain
members evaluation.
Several research questions will be investigated: (1) what are the optimal retail
prices for products sold to different markets? (2) how frequent should the production
process take place? (3) what are the optimal suppliers and components that should be
selected? (4) what is importance of each supply chain member to the supply chain in
an integrated process?

1.5 RESEARCH FRAMEWORK

The overall objective of this research, i.e., coordination of supply chain

15

Chapter 1

configuration decisions for multi-level supply chains, is fulfilled with a conceptual

framework of input-process-output diagram as Figure 1.2. This framework also
provides a general applicable process for CCSC problem, regardless of which
scenario is investigated. This conceptual framework includes three phases (i.e., input,
process, and output), each of which is executed through one or several micro steps.
Explanations for each step are as follows:

INPUT

Domain of CCSC

PROCESS

Scenario
design

Modeling &
Solution

Scenario I

1. Mathematical
modeling

1. Domain of supply
chain network
Scenario II
2. Domain of supply
chain configuration
decision

OUTPUT

Scenario III

Nash game,
Stackelberg game,
etc.

2. Solution
procedure

Scenario IV

Analytic method,
Computational
method,
Optimization
software

3. Domain of supply
chain coordination
approach

Analysis and
interpretation of
result data
Optimal configuration
decisions
Analystic approach
Numerical example
Sensitive analysis

Managerial
guidelines for
decision support

Scenario V

Figure 1.2. Research framework

Input.
To clearly define the domain of CCSC problem makes it easy to incorporate into
the subsequent modeling and analyses. This phase concerns about the clearly defining
domain of CCSC problem which is the combination of three parts: the domains of
supply chain network, supply chain configuration decisions and supply chain
coordination approaches. The scopes of their values have been illustrated in Section
1.3.

16

Chapter 1

Process.
Step 2.1: Scenario design
After identifying the inputs, during this phase, we will select proper coordination
approaches, supply chain configuration decisions and the specific supply chain
networks to formulate the research scenarios for the CCSC problem. As Figure 1.2
shows, the scenario design sector uncovers the relationship of the scenarios studied in
this thesis. The scenario at the start of an arrow set a foundation for the scenario at the
end of an arrow. For example, Scenario I motivates the work from Scenario II to
Scenario IV. Scenario IV inspires us to study Scenario V from a cooperative
perspective. In fact, the first four scenarios are about coordinating supply chain
configuration decisions. The last scenario can be regarded as a configuration model
which determines supply chain configuration decisions in an integrated supply chain.
Step 2.2: Modeling & Solution
In this step, we must develop appropriate mathematical model(s) to determine the
optimal configuration decisions for each scenario. For the first four scenarios,
non-cooperative game models (i.e., Nash game, Stackelberg game) are mainly
employed to study the situations. For the last one, a mix-integer programming model
is set up. Understanding these approaches can help researchers choose right approach
to study their problems.
The development of appropriate solution procedure helps effectively obtain the

17

Chapter 1

optimal configuration decisions of the CCSC problem. The specific solution

procedure for the scenario depends on the nature of mathematical model. Generally,
these models can be solved by analytical method, computational method, or even
some optimization software.
Output.
Step 3.1: Analysis and interpretation of result data.
This step is to identify and conduct necessary analyses for the scenarios and
process and present the results in a convenient ways to generate meaning
interpretations. The results should include the optimal configuration decisions and the
maximal profits for the entire supply chain or the individual members. These analyses
can be conducted by analytic approach or based on simulation approach. For analytic
approach, propositions can be derived by using mathematical theories like Calculus.
Simulation approach is based on results from numerical example and sensitive
analyses.
Step 3.2: Managerial guidelines for decision support
The key issue of this step is to derive and formulate managerial guidelines from
the analytical or numerical results to support the decision-making activities in supply
chain management.

18

Chapter 1

1.6 THESIS ORGANIZATION

This thesis is organized as follows. Chapter 2 presents a literature review of the

theories and findings relevant to this research from several topics, i.e., supply chain
coordination, game theory in supply chain management and supply chain
configuration. Opportunities for further investigations are identified and discussed at
the end of this chapter.
Following the proposed overall research framework, the following Chapter 3 to
Chapter 7 constitute the main body of the research, which correspond to Scenario I to
Scenario V. Chapter 3 studies pricing coordination problem in a three-level supply
chain consisting of a single supplier, a single manufacturer and a single retailer.
Different power structures and channel structures also have significant influence on
the pricing decisions.
Chapter 4 considers the pricing and inventory coordination problem in a two-level
supply chain with one single manufacturer and multiple competing retailers. The
manufacturer purchases different raw materials and produces substitutable products
for different retailers. We develop dynamic non-cooperative game to optimize their
pricing and inventory decisions for the manufacturer and retailers respectively.
Analytical method and solution algorithm are developed to determine the equilibrium
of the game. The proposed game model and solution procedures are illustrated
through a numerical example.
Chapter 5 extends the scenario investigated in Chapter 4 to a multi-level supply
19

Chapter 1

chain consisting of multiple suppliers, a single manufacturer and multiple retailers.

The manufacturer purchases raw materials from outside suppliers to produce
substitutable products to different retailers. The problem is modeled as a three-level
nested Nash game. The model and proposed solution procedure are illustrated through
a simulation experiment and a series of sensitivity analyses conducted to understand
the influence of market, production and raw material related parameters on the
decisions and profits of the supply chain and its constituent members.
Chapter 6 incorporates supplier selection into pricing and inventory coordination
problem. This chapter considers a multi-level supply chain consisting of multiple
alternative suppliers, one manufacturer and multiple retailers. The manufacturer
purchases optional components of certain functionality from his suppliers to produce
a set of products for the retailers in different markets. Nested Nash game model is
used to determine the optimal pricing, inventory and supplier selection decisions. A
numerical study from computer industry has been conducted to understand the
influence of different parameters on the optimal decisions and profits of the supply
chain and the chain members. In addition, the influence of component selection
strategy is also investigated.
Chapter 7 puts forward a mixed-integer programming configuration model for
optimal supplier selection, pricing and inventory decisions for the multi-level supply
chain illustrated in Chapter 6. The suppliers, the manufacturer and the retailers would
like to be involved in an integrated manner to maximize the overall supply chain

20

Chapter 1

system profit. A cooperative game theoretic approach is proposed to evaluate the

marginal contributions of the supply chain members for this supply chain in integrated
process.
Finally, Chapter 8 summarizes the major achievements and findings of this thesis,
addresses the limitations of this thesis, suggests possible directions for future research
work, and in the end summarizes the contributions of this thesis as an overall
conclusion.

21

CHAPTER 2
LITERATURE REVIEW

2.1 INTRODUCTION
The first chapter has introduced supply chain configuration and coordination, and
supply chain network in supply chain management. Game theory is employed to
coordinate supply chain configuration decisions in this research. This chapter attempts
to systematically review the theories, methodologies and finding about the above
topics. Firstly, we will review coordination of supply chain configuration decisions.
Then this review will further to the coordination issues with game-theoretic
approaches. Supply chain configuration problem has also been studied in SCM
separately for the past few decades. We will also cover some configuration models in
SCM in this review. Supply chain network is incorporated into the review of supply
chain coordination and thus, we do not review it apart.
This chapter is organized as follows. Section 2.2 reviews the supply chain
coordination problem, which is mainly conducted from five aspects, i.e., pricing
coordination,

inventory

coordination,

pricing

and

inventory

coordination,

coordination of other supply chain configuration decisions and coordination in

multi-level supply chain. Along the above aspects, Section 2.3 further reviews the
literature using game theory approaches to achieve supply chain coordination on
pricing, inventory and supplier selection, etc. In section 2.4, a literature review on the
supply chain configuration problems and methodologies is given. Finally, discussions
22

Chapter 2

about the most promising research opportunities in this research are presented in
Section 2.5.
2.2 SUPPLY CHAIN COORDINATION
A key issue in SCM is to coordinate the supply chain configuration decisions and
activities of independent supply chain members. Coordination in SCM can result in
improved performance and increased profit for the supply chain system and individual
members (e.g., Jeuland and Shugan 1983, Parlar and Wang 1994, Cortett and de
Groote 2000). For example, some researchers (Curran and Ladd 2000; Sandoe et al.
2001) have argued that the use of Enterprise Resource Planning (ERP) software to
ensure information consistency at various levels of decision-making within a firm.
Forecasting researchers note that bullwhip effect results in increased inventories,
poorer product availability and reduced profits at different levels of supply chain,
which was attributed to the fluctuations in order quantities for a product increase as
they move up an uncoordinated supply chain from the retailer to the wholesaler to the
manufacturer (Chopra and Meindl 2001).
In a typical distribution channel, coordination can be classified as vertical and
horizontal (Chan 2004). Vertical coordination focuses on the issues of why and how
the different channel members should consolidate their activities to achieve better
system performance. Horizontal coordination refers to the coordination of different
functional areas (e.g., marketing, production) of the same firm to achieve better
performance of the whole firm. We will review papers on supply chain coordination

23

Chapter 2

on the following aspects: horizontal coordination and vertical coordination on pricing

and inventory decisions, some other supply chain configuration decisions (e.g.,
advertising, lead-timing) and coordination of multi-level supply chain.
2.2.1 Coordination of pricing and inventory decisions
2.2.1.1 Price coordination
Many retail and manufacturing companies develop innovative pricing strategies,
such as dynamic pricing, target pricing, to improve their operations. For example, the
largest online retailer, Amazon.com, allows its enterprise clients to quickly and easily
change the prices based on demand variation, inventory level, etc. The benefits
include not only potential increase in profit, but also improvements in demand or
production variability, thus more efficient supply chains (Chan et al. 2004).
Pricing, as a fundamental decision for supply chain management, has been studied
in many marketing research literature on distribution channels. Early works on pricing
often focus on a bilateral monopoly (i.e., a single manufacturer sells through an
exclusive but independent retailer) model and have identified vertical coordination
through quantity-discounts. Jeuland and Shugan (1983) study a simple system where
one manufacturer sells items to a single retailer. They show that quantity discounts
can provide an optimal means for achieving coordination. They also show that the
manufacturer can obtain higher profit and increase profits for the entire channel
simultaneously by horizontal coordination. Over the years, numerous researchers
(McGuire and Staelin 1983; Moorthy 1987; Jeuland and Shugan 1988) have

24

Chapter 2

confirmed that Jeuland and Shugans (1983) findings about the advantages of channel
coordination are correct only if there is no product competition. Moorthy (1987)
proposes a simpler coordination method by using a two-part tariff and showed that a
quantity discount scheme is not necessary for channel coordination. Jeuland and
Shugan (1988) point that under certain circumstance rational conjectures can lead
profit maximizing channel members to attain some degree of coordination.
Monahan (1984) is the first to illustrate how a monopolistic manufacturer
operating as a leader can increase its profit by providing a discount. The
manufacturer tries to induce larger order quantities from the retailer to reduce its own
order processing cost. To compensate for the increase in holding cost due to larger
order size, the retailer receives a price discount. Overall, the retailer is no worse off
and the manufacturer reaps all the benefits. Lee and Rosenblatt (1986) generalize
Monahans model by relaxing the lot-for lot production assumption of the
manufacturer. They develop an algorithm to determine quantity discount policy for
the manufacturer to the retailer, which maximize the manufacturers profit.
By and large, these researches on coordination for pricing decision ignore
inventory information of a product in modeling the supply chain profit models. Lal
and Staelin (1984) analyze a single seller and a single buyer channel and show it is
possible for the seller design a pricing policy where the average price of the sellers
product is a monotonically decreasing function of the amount ordered by the buyer
and the sellers profit is increased while the buyers profit would not be reduced.

25

Chapter 2

Choi (1991) incorporates channel and products competition into pricing models.
Choi examines a pricing problem for a channel with two competing manufacturers
and one common retailer who sells both manufacturers products. The author also
studies the effects of power structure, product differentiation, and cost differences on
channel coordination and finds that demand function employed significantly
influences the results. Choi (1996) extends Chois (1991) analysis by including price
competition between duopoly common retailers. His work not only captures product
differentiation, but also models store differentiation between the retailers. Choi shows
that while product differentiation benefits manufacturer, it hurts retailers. In the
contrary, while store differentiation benefits retailers, it may hurt manufacturer.
Ingene and Parry (1995a, b; 1998; 2000) explore the channel coordination by a
manufacturer that sells an identical product to multiple competing independent
retailers. They derive a linear quantity discount schedule which duplicates the retail
price and quantity results of a vertical integrated system. They show that a
Stackelberg two-part tariff that could not coordinate the channel and a
second-best two-part tariff that approximates the results from an integrated channel.
Then, the authors show that in some cases, the manufacture would rather offer either
the Stackelberg tariff or the second-best tariff over a channel-coordinating
quantity-discount schedule.
Gerstner and Hess (1995), as a departure from offering incentives to retailers,
recommend the manufacturer to use a pull price promotion directly to price-sensitive

26

Chapter 2

consumers as a coordination device in an independent channel of distribution. The

price coordination improves channel profits and consumer surplus.
2.2.1.2 Inventory coordination
Early researchers have addressed inventory coordination issues in a single supplier
and single buyer channel. Goyal (1977) suggests a joint economic lot size model to
minimize the total relevant costs for both the vendor and the buyer with an infinite
production rate for the vendor. Banerjee (1986) extends Goyal (1977) model by
considering a finite production rate for the vendor to obtain the optima joint
production or order quantity. Goyal (1988) further generalizes Banerjees model by
assuming that the vendors lot size is an integer multiple of the buyers order size.
Goyal shows that his model provides a lower or equal joint total relevant cost
comparing Banerjee (1986). However, this model is based on the assumption that the
vendor can supply the buyer only after completing the entire lot size.
Lu (1995) relaxes the assumption in Goyals (1988) model and allows that the
vendor can afford multiple buyers even before completing the entire lot size. Lu
proposes the optimal solution procedure for one-vendor one buyer case and suggests a
heuristic approach for one-vendor multi-buyer case. Goyal (1995) uses the numerical
example in Lu (1995) to show a different shipment policy can give a better solution.
Hill (1997) deals with a single vendor and a single buyer channel and proposes a more
general class of shipment policy.
VMI (vendor managed inventory) is a new practice for improving the inventory

27

Chapter 2

management in supply chains. In a VMI system, a vendor holds large quantities of

products and decides on the appropriate inventory levels of each of the products for
all his retailers, and the appropriate inventory policies to maintain these levels
(Simchi-Livi et al. 2000). Banerjee and Banerjee (1992) consider a VMI system
where the vendor makes all replenishment decisions for his buyers to improve the
joint inventory cost. Woo et al. (2001) extend Banerjee and Banerjees (1992) work to
incorporate ordering cost reduction and raw material purchasing into inventory system.
They investigate an integrated inventory system in which a single vendor purchases
raw materials and produces finished items to multiple buyers. Dong and Xu (2002)
study the effects of VMI and conclude that VMI can realize many benefits which is
obtainable only in fully integrated supply chain. Since then, VMI is widely studied by
other researchers (Disney and Towill 2003; De Toni and Zamolo 2005; Rusdiansyah
and Tsao 2005; Yao et al. 2007). Yu et al. (2009a, b) consider a VMI supply chain
with one manufacturer and multiple retailers. They extend their two-echelon
inventory supply chains to three echelon ones including his raw materials inventory
for the manufacturer.
2.2.1.3 Pricing and inventory coordination
For many manufacturing industries, the coordination of pricing with inventory and
production decisions, means an approach that improves the efficiency of both the
supply chain and individual firms. Table 2.1 summarizes and categorizes the literature
reviewed in this section.

28

Chapter 2

Table 2.1 References for pricing and inventory coordination

Pricing inventory
coordination

Characteristic

Reference
Whitin 1955; Kunreuther and Richard 1971;

Coordination within a firm

Tersine and Price 1981; Arcelus and Srinivasan

1987, Ardalan 1991,

Horizontal
Incorporating suppliers pricing

1990; Lee 1993; Abad 1997,

strategy
One supplier and one buyer (or
homogeneous buyers)

Vertical

Lal and Staelin 1984; Abad 1988; Burwell et al.

Weng and Wong 1993; Parlar and Wang 1994;

Weng 1995a, 1995b

One supplier and

Federgruen and Heching 2002; Boyaci and

geographically dispersed buyers

Multiple products or multiple

Gallego 2002
Cheng 1990; Chen and Min 1994b; Morgan et al.
2001; Kachani and Perakis 2002,

customers

Many early researches focus on horizontal coordination on pricing and inventory

issues. Whitin (1955) is the first to add pricing decisions into inventory problems. He
studies the problem faced by a single decision-maker who orders a product from a
supplier that charges a constant unit cost. The demand for the firm is deterministic
linear function of price and the objective is to maximize expected profit. This study
shows that the retailer who coordinates the price and order quantity decisions could
obtain greater profits than one that allows two separate departments (marketing and
operations) to make these two decisions in an uncoordinated manner. Kunreuther and
Richard (1971) analyze the same model, but focus on the situation that the marketing
department sets price and purchasing department determines the order quantity. They
consider the difference between the centralized solution and a sequential solution,
where the marketing department assumes the unit cost to be the sum of the different
variable cost and the purchasing department uses the resultant demand rate to

29

Chapter 2

determine the order quantity. They show in the numerical example that the centralized
solution produces a higher profit with a higher price and smaller lot size. They also
extend Whitins model to a finite production rate case and find that optimal lot size is
larger and optimal price is smaller than in EOQ framework. Over the years, Whitins
and Kunreuther and Richards work has been a foundation to a steam of research
(Tersine and Price 1981, Arcelus and Srinivasan 1987, Ardalan 1991, Hall 1992,
Martin 1994, Arcelus and Srinvasan 1998) which draw the same conclusions in
various circumstances.
There is a closely related line of research that incorporates the suppliers pricing
strategy into a firms optimal pricing and inventory decision-making. These pricing
strategies offered by the supplier are often regarded as input parameters. Lal and
Staelin (1984) show it is possible for the seller design a pricing policy where the
average price of the sellers product is a monotonically decreasing function of the
amount ordered by the buyer and the sellers profit is increased while the buyers
profit would not be reduced. Abad (1988) concerns with optimizing price and lot size
for a retailer, who purchases all-unit quantity discounts product from the supplier. The
author presents a procedure using both centralized and decentralized approaches.
Burwell et al. (1990) find Abads procedure performed near-optimally for the case of
linear demand. Lee (1993) further indicates that Abads procedure may result in
suboptimal solutions and proposes an optimal procedure based on geometric
programming. Abad (1997) considers the case in product distribution that the supplier

30

Chapter 2

offers a temporary reduction in price to engage in forward buying. The buyer faces the
price sensitive demand and determines both the selling price and order quantity to
maximize his profit. A similar scenario is studied by Ardalan (1994, 1995). The author
uses both average cost and present value of cost methods to develop inventory
policies and determine optimal policies when there is a temporary change in price and
demand.
More recently, vertical coordination has received a good deal of attention in the
supply chain context. These studies mainly focus on coordination issues on pricing
and order quantity decisions between suppliers (sellers) and retailers (buyers) in
decentralized multi-stage inventory systems. Weng and Wong (1993) investigate the
effects of quantity discount policies under various conditions. In their model, multiple
types of setup costs are included: the buyers order setup cost, the suppliers outbound
order handling and order processing cost. They also consider the ordering decision for
the supplier the integer ratio of his purchase quantity to the buyers lot size. Parlar
and Wang (1994) study a special case of the problem using linear demand function
and a constant multiple replenish policy. They show that the supplier might benefit
from the quantity discount policy if the buyer is provided sufficient incentive to
increase his order quantity from the status quo. Weng (1995a, b) investigates the
effects of joint decision policies, i.e., optimal joint selling price and order quantity, a
channel coordination problem in a system consisting of a supplier and a group of
homogeneous buyers. It is shown that Monahans (1984) optimal all unit quantity

31

Chapter 2

discount policy was equivalent to Lal and Staelins (1984) optimal incremental
quantity discount policy.
The work discussed above pertain to one supplier and one buyer (or homogeneous
buyers) channel. Federgruen and Heching (2002) study joint pricing and inventory
problem in distribution systems geographically dispersed retailers. A distribution
center places an order with an outside supplier and then allocates to the retailers. They
develop an approximate tractable model, which provides analytically computable
approximations for system profits and other performance measure. They also develop
combined pricing, ordering and allocation strategies and show that the systems
performance under these strategies is well gauged by the above approximations.
Boyaci and Gallego (2002) analyze the problem of coordinating pricing and inventory
replenishment policies in a supply chain consisting of a wholesaler, one or more
geographically dispersed retailers. The market demand is a deterministic
price-sensitive function. The authors show that optimally coordinated policy could be
implemented as a nested replenishment policy where the retailer pays the wholesale
the unit wholesale price as the items are soled, and the retail price is set jointly with
the manufacturer. They also show pricing and lot sizing decisions can be made
sequentially without loss of optimality when demand is large.
We then discuss papers that treat multiple products or multiple customers. Cheng
(1990) proposes an economic order quantity (EOQ) model that integrates pricing and
order sizing decisions for multi-product problem. The author uses the Kuhn-Tucker

32

Chapter 2

conditions to determine the optimal pricing and order sizing decisions with the
assumption of the same order frequency and storage capacity limitations. Chen and
Min (1994) indicate that the solutions provided by Cheng may result in non-optimal
solution under linear demand assumption and also derive an optimal closed-form
solution and the conditions to guarantee optimality. Morgan et al. (2001) present a
model for a mix of product selection incorporate marketing inputs with manufacturing
cost information. However, the prices of products are given as inputs. Kachani and
Perakis (2002) investigate pricing and inventory problem with multiple products
sharing resources and apply fluid method to determining the optimal pricing,
production and inventory decisions. Zhu and Thonemann (2002) consider a pricing
and inventory coordination problem for a retailer selling two products. The demand of
each product depends on both products prices. Their studies show that the optimal
pricing and inventory policies and show that the retailer can benefit if both products
are managed jointly.
2.2.2 Coordination on other supply chain configuration decisions
Except for pricing and inventory coordination, in recent years, coordination of
some other supply chain configuration decisions has also attracted much attention. In
this section, we will brief review the literature concerned with the coordination of
such supply chain configuration decisions, as advertising, lead time, supplier
selection.
Advertising, apart from its numerous direct effects, also has an effect on the price

33

Chapter 2

sensitivity of consumers and on the prices of goods in market. Managers and

researchers are interested in coordinating advertising and pricing decisions to attain
maximum profit. Bergen and John (1997) consider wholesale price, retail price,
participation rate and intra-brand competition among multiple retailers. Their model
focuses on local advertising to study the most prominent aspect of these plans the
effect of manufacturer participation. Huang et al. (2002) explore the role of
cooperative (co-op) advertising in a manufacturer-retailer supply chain by brand name
investments, local advertising expenditures, and sharing rules of advertising expenses.
Yue et al. (2006) extend the co-advertising model of Huang et al. (2002) by
incorporating price elasticity to study the effect of direct manufacturer price discount
on channel coordination. Karray and Zaccour (2006) find that co-advertising can be
an efficient counter-strategy for the manufacturer when a retailer offers a private label
product that may affect the national brand.
Except for advertising, lead time is also another supply chain configuration
decision involved in supply chain coordination. The importance of product delivery
lead time to the competitiveness of a firm has been recognized in practice and
literature for some time. Yano (1987a, b, c) is one of the first authors to study how to
determine planned lead times for a fixed batch size in a two-stage serial/ distribution
system and two components in a two-stage assembly system. Hopp and Spearman
(1993) consider how to set safety delivery lead times for multiple components with
random procurement times and deterministic assembly times. Liu et al. (2007) study

34

Chapter 2

pricing and lead time decisions of a decentralized supply chain consisting of a

supplier and retailer.
In the last few decades, the supplier selection problem has gained great attention
in business literature and practices. Webber and Current (1993) define the supplier
selection problem as two joint decisions- which suppliers should be selected and how
much should be ordered from the selected suppliers. The supplier selection is
necessary to be integrated in the context of supply chain management. However, most
of the supplier selection literature do not consider other supply chain configuration
decisions (de Boer et al. 2001).
Degraeve and Roodhooft (1999) propose a mathematical programming model
using the total cost of ownership (TCO) to simultaneously select suppliers and to
determine the order quantities over a multi-period time horizon. Degraeve and
Roodhooft (2000) further develop a multi-period, multi-item, multi-vendor mixed
integer programming model based on the TCO, to determine an optimal ordering and
inventory police and jointly to decide the best combination of suppliers. Their model
covers the total cost incurred, including the purchasing cost, the ordering cost, the
transportation costs and so forth. Ghodsypour and OBrien (2001) build
single-objective and multi-objective mathematical programming models minimizing
the total cost of logistics in the process of the supplier selection, including the
aggregate price, the ordering cost and the inventory costs, subject to capacity, budget,
quality and delivery requirements. Their models apply the economic ordering quantity

35

Chapter 2

(EOQ) model to determine the lot sizing decision. Liao and Rittscher (2007) establish
a multi-objective programming model, integrated supplier selection, procurement lot
sizing and carrier selection decision for a single purchasing item over multiple
planning periods. In these research, the authors efficiently incorporate supplier
selection decisions with inventory polices to improve the channel profits. However,
marketing strategies are not always involved in their models.
2.2.3 Coordination of multi-level supply chains
Previous studies on supply chain coordination mainly examine supply chains
consisting of only two levels. Some researchers turn their attention to coordination
issues in multi-level supply chains.
Alan

and

Medini

(1992)

study

pricing

in

three-level

system

(manufacturer-retailer-customer). They conclude that, in order to maximize the profit,

the manufacturer would cooperate with the retailer to sell the product to the customer.
Munson and Rosenblatt (2001) explore quantity discount mechanism to coordinate
inventory in a three-level supply chain (supplier-manufacturer-retailer). They show
that incorporating quantity discounts into both ends of the supply chain can
significantly decrease costs compared to concentrating only the lower end.
Moutaz (2003) investigates three inventory coordination mechanisms in a supply
chain with multiple firms at each stage and a firm can supply two or more customers.
The results show that some of the coordination mechanism can lead to significantly
lower total cost than matching production and delivery along the chain.

36

Chapter 2

Giannoccaro and Pontrandolfo (2004) propose a contract model based on the

revenue sharing mechanism to coordinate a decentralized control three-stage supply
chain. The authors consider two contract models: one is offered by the manufacturer
to the distributor, the other one is offered by the distributor to the retailer. They show
that an integrated design of the two contracts lets the retailer and the distributor select
order quantities that are optimal for the whole supply chain.
Jabber and Goyal (2008) study the coordination of order quantities between the
members in a multiple suppliers, a single vendor and multiple buyers supply chain.
The coordination model achieves coordination amongst the members in a supply
chain assuming common cycle time for all non-identical buyers.
These above studies focus on coordination of pricing or inventory policies in
multi-level supply chains with one or multiple firms at each stage. However, none of
them incorporates multiple elements of supply chain configuration (e.g., both pricing
and inventory) in the supply chain coordination issues.
2.3 GAME THEORY IN SCM PROBLEMS
Many economists in the past century have work on how to use game theory in
analysis of supply chain with multiple agents, often with conflicting objectives. The
classic book Theory of Games and Economic Behavior written by von Neumann
and Morgenstern (1944) summarizes the basic concepts existing at that time. From
then, game theory has gained enormous development in many aspects, including the
concept of equilibrium by Nash (1950), games with imperfect information by Kuhn

37

Chapter 2

(1953), cooperative games by Aumann (1959), etc.

Generally, a game consists of (1) players indexed by n=1, 2,, N, (2) strategies
or more generally a set of strategies denoted by xn , n=1, 2, , N, for each player, (3)
payoffs n x1 , x2 ,..., xn , n=1, 2,, N, received by each player. Each strategy belongs
to a strategy set X n , xn X n . The Cartesian product X1 X 2 ... X n is the
strategy profile set. In most SCM problems, players are always regarded to be the
supply chain members (e.g., Lal and Staelin 1984, Wang 2002). A players strategy
can be considered as the complete instruction for the actions to take in a game
(Cachon and Netessine 2005).
This section focuses on the literature that are particularly relevant to the
coordination of supply chain configuration decisions through non-cooperative and
cooperative game-theoretic approaches. Table 2.2 presents a brief summary and some
examples of the literature of supply chain coordination game model in two categories
non-cooperative and cooperative. The literature in each category will be further
reviewed based on three different types of models pricing game, Inventory cost
game, and pricing and inventory game.
2.3.1 Pricing game model
2.3.1.1 Pricing game model
When competing channel systems are considered, complexity of pricing emerges
as the suppliers, retailers, and product variety increase. Complicated interactions
among different channels and channel members are extensively examined by game

38

Chapter 2

theory.
Table 2.2. Game models for supply chain coordination
Game type

Model

Key problem

Non-cooperative

Cooperative

Devise coordination mechanism

Divide

that can coordinate the activities

profits that have been generated

and

once cooperative relationship

align

the

independent

objectives
supply

of

chain

the

additional

total

achieves (Driessen 1988).

members (Chen et al. 2004).

Pricing game
model
Inventory cost
game model
Pricing and
inventory game
model

Monahan, 1984; Lal and Staelin,

Dada

1984;

Chakravarty and Martin, 1988;

Choi, 1991;

Minakshi,

1998
Lariviere

and

Srikanth,

1987,

Kohli and Park, 1989

and

Porteus

2001;

Meca et al., 2004

Slikker et al., 2005; Rhim et al.,

2003; Wang and Gerchak, 2003
Bernstein and Federgruen, 2003;

Abad, 2003;

Chen et al., 2004; Huang et al.,

al.,2008

Esmaeili et

2007

McGuire and Staelin (1983) investigate the effect of product substitutability on

the pricing decision in a distribution channel where each manufacturer distributes its
products through a single retailer. They find that for low degrees of substitutability,
each manufacturer will distribute its product through a company store; for highly
substitutive products, then decentralization is a Nash equilibrium strategy for the
manufacturers. Moreover, this equilibrium is preferred by both manufacturers to the
other equilibrium when both manufacturers vertically integrate. Moorthy (1988)
further explains the effect of strategic interaction on channel structure by identifying
the necessary conditions for decentralization to be a Nash equilibrium strategy, which
is a supplement to the examples in McGuire and Staelin (1983).
Choi (1991; 1996) examines the product and store differentiation in channel with

39

Chapter 2

two competing manufacturers and one / duopoly common retailers who sells both
manufacturers products. Choi studies three non-cooperative games of different power
structures between the two manufacturers and the retailer, i.e., two Stackelberg and
one Nash games. His studies show that the price competition between differentiated
products or stores plays a critical role in determining the channel structure. Minakshi
(1998) extends the literature by analyzing three channel structures, dealing with two
competing manufacturers and two retailers. The author considers Stackelberg game
structure between the manufacturer and the retailer and shows that the introduction of
competition at retail level adds a new dimension to the game which has a significant
impact on channel profitability and choice.
2.3.1.2 Inventory cost game
Inventory cost game is introduced in Meca et al. (1999) and Meca (2000). In this
type of game, inventory models have been approached from the perspective of the
cooperation and competition of the agents involved. The agent players deal with the
ordering and holding of a certain commodity and decide to make their orders jointly
(Meca 2003).
In particular, the newsvendor problem selling product with short lifecycle to
customers, has been widely studied in Hartman et al. (2000), Lariviere and Porteus
(2001), and Slikker et al. (2005). Talluri and Baker (2002) propose a three-phase
mathematical programming approach for efficient supply chain design in a three-level
supply chain of suppliers, manufacturers and distributors. In the first phase, they use

40

Chapter 2

pair-wise efficiency game techniques to evaluate the performance of supply-chain

members. In the second phase, they apply an integer programming model to design
supply chain to efficiency, capacity and location constraints. In the last phase, they
identify the overall optimal routing decisions.
Rhim et al. (2003) formulate a supply chain design problem as a three-stage
non-cooperative game covering location, capacity, and production quantity decisions.
Nash equilibrium is used as the solution concept in each stage. Wang and Gerchak
(2003) analyze a capacity-building problem in a multi-stage production network
including multiple sellers and a single buyer as Stackelberg games. They consider two
scenarios in which the sellers and the buyer act as the leader respectively and compare
the results to find the situations where a game mechanism is proper.
2.3.1.3 Pricing and inventory game model
The interaction between pricing and production/distribution efficiency has been
studied by a number of researchers. Next, we will review the ones that include
competition under price and inventory coordination using non-cooperative game
theoretic approaches.
Bernstein and Federgruen (2003) study a two-echelon distribution channel with
one supplier serving a network of competing retailers. They compare the optimal
performance of the centralized supply chain with that of various decentralized supply
chain and distinguish systematically between the case of Bertrand and Cournot
competition. However, Bernstein and Federgruen mainly focus on the competition

41

Chapter 2

between the retailers.

Chen et al. (2004) consider a horizontal market of multiple firms who make joint
pricing and inventory decisions and use price to compete for market demand. They
prove the existence and uniqueness of the pure-strategy Nash equilibrium and show
that the market at the equilibrium exhibits a bias toward under-pricing caused by
competition. Parlar and Weng (2006) joint consider the pricing decision faced by the
marketing department and the production quantity decision faced by the production
department. They allow the two departments to coordinate their decisions to compete
against another firm. They analyze two scenarios by game-theoretic approaches, i.e.,
no coordination and coordination in both firms. They show that the coordination of
pricing and production quantity decisions in a firm could lead to a stronger
competitive position in terms of lower prices, larger production quantity and higher
expected profit.
Huang et al. (2007) incorporate supplier selection and product design into pricing
and inventory coordination problem in a supplier-manufacturer supply chain. They
formulate a three-move non-cooperative dynamic game to solve this problem. In the
first move, the manufacturer makes decision on the component and supplier selection
decisions. Then the selected suppliers make the second move to optimize their
decisions including price discounts and ordering policies. The manufacturer will
finish the game by taking the third move to make his ordering decisions. Yu et al.
(2009b) consider optimal pricing, inventory policies and advertising problem in a

42

Chapter 2

vendor managed inventory (VMI) supply chain with one manufacturer and multiple
retailers. They formulate this problem as a Stackelberg game where the manufacturer
is the leader and the retailers are the followers.
The game model reviewed above mainly focus on pricing and inventory decisions
for the competing firms in either one-echelon or two-echelon distribution system. The
effects of coordination are examined by different non-cooperative game approaches.
2.3.2 Cooperative game
Profit sharing mechanisms have also been proposed for supply chain coordination.
Under this mechanism, the performance of the supply chain system is first optimized
and the resultant benefit is then shared between the supply chain members.
Cooperative game is employed as the profit sharing scheme.
Some studies on pricing problems have been modeled as integrated seller-buyer
cooperative game where the seller and the buyer would be involved in a win-win
pricing policy cooperatively and split the overall savings in some negotiation or
bargaining. The examples can be seen from Dada and Srikanth (1987), Chakravarty
and Martin (1988), Kohli and Park (1989), Kim and Hwang (1989). In these studies,
complete cooperation is the essential assumption between the players to maximize the
system-wide payoffs.
Other related studies pricing problem by comparing of different game-theoretic
approaches. Chiang et al. (1994) investigate the quantity discount pricing problem of
a single supplier and a single buyer channel using non-cooperative game and

43

Chapter 2

cooperative game. Stackelberg equilibrium and Pareto Optimality criteria are used as
the solutions of non-cooperative game and cooperative game. Through the
comparison between the two games, they show that both the vendor and buyer could
be better off if they can cooperate. Yang and Wee (2003) study the system integration
and quantity discount problem of a seller-buyer inventory system under three
coordination scenarios. In the first one, the buyer optimized his own costs and
imposes the resulting policy to the vendor. In the second scenario, the vendor and the
buyer are totally integrated without quantity discount policy. In the last scenario, the
buyer and the seller work cooperatively to minimize system wide cost.
Except for pricing problems, cooperative game-theoretic approaches have also
been employed to study pricing, inventory and some other supply chain configuration
problems. For example, Abad (2003) also determines pricing, lot size and trade credit
policies in a seller-buyer channel by non-cooperative and cooperative game
approaches. The demand is price sensitive and the seller may offer trade credit to the
buyer. Their work provides a framework to the seller for coordinating his pricing and
credit policy. Esmaeili et al. (2009) propose several game models of seller-buyer
relationship to optimize pricing and lot sizing decisions. Meca et al. (2004) study how
cooperation can minimize the joint inventory cost of a decentralized multi-stage
supply chain network through a cooperative game-theoretic approach and how to
share the joint cost among the game players. Smirnov et al. (2004) formulate the
supply chain configuration problem as a coalition game which is solved by a

44

Chapter 2

combination of soft-computing techniques.

2.4 SUPPLY CHAIN CONFIGURATION
In the past few decades, methodologies and applications of supply chain
configuration become an important issue of the SCM. Most supply chain
configuration problems are formulated as centralized models. Mixed-integer
programming is the commonly used and appropriate modeling approach for
centralized supply chain configuration problems (Shapiro 2001). Geoffrion and
Graves (1974) are among the early researchers who use mixed integer programming
to solve such problem. They develop a multi-commodity logistics network model for
optimizing product flows from plants to distribution facilities and to final customers
and a solution technique based on Benders decomposition. Cohen and Lee (1988;
1989) formulate a deterministic, mixed integer optimization program based on
economic order quantity policy to consider for global manufacturing and distribution
networks design. Lee and Billington (1995) apply these models to validate the global
manufacturing strategies of Hewlett-Packard. Arntzen et al. (1995) present a
comprehensive deterministic mixed integer program model for designing global
supply chain to determine the optimal manufacturing and distribution strategies to
minimize total cost, weighted cumulative production and distribution times. The
model was successfully applied in the Digital Equipment Corporation. Munson and
Rosenblatt (1997) investigate a global supply chain problem using mixed integer
program to select suppliers and final production sites, and allocate purchase quantities

45

Chapter 2

for a particular market to minimize the sum of purchasing, production, transportation,

and fixed costs. Syarif et al. (2002) propose a mixed-integer programming model to
formulate a logistic network problem to design distribution network.
Despite of extensive application to model supply chain configuration problems,
mixed-integer programming models are usually hard to solve for the huge number of
feasible decisions. Thus, some optimization software, such as Lingo, is always
adopted to assist solving these models. Dominguez and Lashkari (2004) consider
capacitated, multi-stage, multi-period, multi-commodity, multi-facility inventory
planning problem in the context of the real-world, large-scale business operations of
an appliance manufacturer, AG, inc., Mexico. They develop a complex mixed-integer
programming model and solve it by Lingo. Hamedi et al. (2009) investigate a
real-world case in a natural gas supply chain. Their study uses a mixed-integer
non-linear programming model to minimize direct or indirect distribution costs. This
model is also solved by Lingo.
Recently, there has been a growing interest in using hybrid models to configure
supply chains. Genetic algorithms (GA) and some other algorithms are the most
adopted methods for solving these models. Byrne and Bakir (1999) develop a hybrid
algorithm combining mathematical programming and simulation models for a
multi-period, multi-product production planning in a manufacturing system. Lee and
Kim (2000; 2002) combine analytic and simulation models to deal with the similar
problem. They show that the proposed hybrid approach could provide better solutions

46

Chapter 2

when there exist stochastic characteristics. Talluri and Baker (2002) propose a
three-phase mathematical programming approach for supply chain design. Phase one
is to evaluate the performance of potential suppliers, manufacturers, and distributors.
Phase two uses an integer programming model to select candidates for supply chain
configuration. Phase three determines the optimal trans-shipment, deployment and
routing decisions. Kim et al. (2002) is among the few ones, who develop a supplier
selection optimization model for a multi-commodity production network composing
of a manufacturer and its suppliers. They also present an iterative solution algorithm
to determine the ordering decisions for this multiple sourcing problem.
Graves and Willems (2005) develop supply chain configuration model for a
multi-stage production/distribution network with normally distributed demands to
minimize the total supply chain cost and demonstrate the success by applying to a
notebook computer manufacturing supply chain. Huang et al. (2005b) solve the same
problem by GA, achieving higher computational efficiency and compare supply chain
configurations with different levels of parts commonality. Truong and Azadivar (2005)
concurrently consider a variety of decisions associated with supply chain management
in general and supply chain configuration in particular by the combination of mixed
integer

programming

and

GA.

Ko

et

al.

(2006)

propose

hybrid

optimization/simulation approach to configure a distribution network from the

perspective of third-party logistics provider. The optimization model use GA to
determine dynamic distribution network structure, and then simulation model is used

47

Chapter 2

to evaluate the configuration with regard to demand and transportation uncertainty.

2.5 SUMMARY
This chapter has revealed the accumulation of a rich body of literature in the field
of supply chain coordination and configuration over the past decades. The related
literature can be divided into three categories. The first category includes studies
related to supply chain coordination covering pricing coordination, inventory
coordination, etc. A great number of these research focus one or two aspects of supply
chain configuration, e.g., pricing, inventory, and two-echelon supply chain network,
e.g., supplier-retailer channel. The second category involves the literature employing
non-cooperative and cooperative game theoretic approaches to study the supply chain
coordination problem. The third category includes studies related to supply chain
configuration problems where the major modeling approaches as well as the solution
algorithm and software are discussed. Based on the review of literature presented in
this chapter, we come up with the following conclusions that inspire our research.
Pricing and inventory are two important supply chain configuration decisions.
Coordinating both of the decision has been proved of much benefit to the supply chain
system and the individual chain members. However, studies in supply chain
coordination issues on pricing decision, inventory decision, or both have mainly
limited in two-echelon distribution system, e.g., supplier-retailer relationships. The
coordination of either or both decisions in multi-echelon supply chain is very rare.
Inventories exist throughout the supply chain in various forms. Managing

48

Chapter 2

inventory cost is a major challenge in supply chain management. In the literature for
two-echelon supply chain, such as manufacturer-retailer relationship, few of them
consider the manufacturers raw materials inventory. However, the manufacturer is
necessary to keep raw material stock to ensure continuity of production. Moreover,
how to formulate the inventory for the upstream suppliers of the manufacturer will be
another concern of our research.
Most of the existing literature in supply chain management focus on competition
between echelons of the supply chain or the individual firms in the same echelon. For
the competition in single echelon, such as Bernstein and Federgruen (2003), mainly
consider the competition between different retailers. The interaction between
upstream suppliers is always not involved in their work. Some other research may
focus on the competition between different echelons, while omitting the competition
between the members in the same echelon, e.g., Yu et al. (2009). The studies
considering both types of competition have not received adequate attention.
Analytical method has been an important approach to analyze supply chain
coordination problems, especially the equilibriums of different game models.
However, the analytical method is limited to simple supply chain networks, e.g., the
supply chains with one member in each echelon. Computational approaches can help
solve the complex supply chain problems. Recently, the combination of analytical and
computational methods has been preferred to solve complex supply chain
management research.

49

Chapter 2

Typical supply chain configuration decision supplier selection has also been
widely studied separately. Recently, it is widely acknowledged the advantage of
integrating supplier selection with other supply chain configuration decisions, e.g.,
pricing, inventory decision. However, the coordination of supplier selection and other
supply chain configuration decisions is sparse, especially for multi-echelon supply
chains. Opportunities exist in this area for further rigorous research.
Cooperative game is widely employed to study the profit division between the
chain members, which are according to their negotiate power or contribution. Because
of its complexity, its use to study multi-level supply chain is limited. How to
formulate cooperative game in multi-level supply chain and use it to study the
properties of the supply chains is another area deserved further study.

50

CHAPTER 3
GAME-THEORETIC PRICING COORDINATION IN A MULTI-LEVEL
SUPPLY CHAIN WITH DIFFERENT CHANNEL STRUCTURES
3.1 INTRODUCTION
Pricing is an important supply chain configuration decision affecting the
profitability of an enterprise and also plays a significant role in balancing demand and
supply (Alan and Medini 1992; Elmaghraby and Keskinocak 2003). Pricing strategy
in the supply chain wide is not only concerned with enterprises individually, but other
chain members mutually, as well as the entire supply chain system. However, chain
members often have conflicting objectives among themselves. Pareto-optimal pricing
decisions cannot always be achieved (Choi 1991). Hence, coordination of different
members of the channel must be considered (Martha and Lisa 1993).
Different power structures and channel options affect the pricing decisions for the
supply chain members differently (Choi 1991; Chen et al. 2006). Selecting
appropriate power structure and channel structure for distribution is a major concern
for each firm of the supply chain. In many supply chains, some powerful
manufacturers traditionally hold manipulative powers, act as the channel leader, and
are followed by other chain members. For example, a powerful manufacturer, such as
Proctor & Gamble (P&G), may be able to demand certain shelf spaces in a retailers
stores (Burns 1994). McDonalds typically imposes strong quality control over its food
suppliers (Brookes 1995). Rangan et al. (1992) highlight the importance of channel

51

Chapter 3

selection. In many industries, the manufacturers forward integration with its

downstream retailer would like to work together with its downstream retailer in
determining pricing, advertising, etc., to cut down the cost, lower the retail price
(Fisher, 2003). For example, P&G now cooperates with Wal-Mart through JIT
delivery, information sharing, and demand monitoring, etc (Coyle et al. 1996; Foley
and Mahmood 1996). Other examples of such systems are Hart, Marx (operating more
than 275 stores), International Harvester, and Sohio, etc. Effective integration of
suppliers into supply chain will be a key factor for the manufacturer to remain
competitive (Robert, et al. 1999). Sears, like many other large retailers, owns about
one-third of its supplier DeSotos (a furniture and home furnishings manufacturer)
stock. Some enterprises would also like to achieve full vertical integration - forward
and backward integration. The illustrations can be found in the airline industry, oil
refining companies, etc. Oil refining companies have traditionally owned their oil
distribution channels, such as gas station. Sometimes, they also move into oil
exploration and exploitation. Under the different power structures and integration
channels, the pricing decisions of the chain members will also be different.
This chapter conducts an investigation into a single product three-level price
models composed of one supplier, one manufacturer and one retailer with price
sensitive demand. Three different channel structures are considered in this supply
chain. The first is the decentralized channel that the manufacturer uses independent
suppliers and retailers, in which they optimize their own profit individually and

52

Chapter 3

non-cooperatively. The second channel is that the manufacturer backward integrates

with the supplier or forward integrates with the retailer and uses independent retailer
or supplier simultaneously. This channel is called by semi-integrated channel, because
it is partially integrated to the overall supply chain. The third one is the integrated
channel structure. In this channel, the manufacturer, the supplier and the retailer
operate in total cooperation in a full vertical integration system. Two power balance
scenarios are considered for the decentralized channel and the semi-integrated channel.
The first leader-follower power structure, where the manufacturer always takes the
leadership, is modeled as a Stackelberg game in which the leader takes the followers
reaction function into consideration for its first move on pricing decision; the second
independent power structure is treated as a simultaneous non-cooperative game (i.e.,
Nash game) where each members are of equal status and they make their pricing
decisions simultaneously and non-cooperatively (Choi 1991).
Through the above supply chain model, we are particularly interested in the
following questions:
(Q3.1) how will the supply chain members of different echelons make their optimal
pricing decisions under different power structures and channel structures?
(Q3.2) which power structure will be preferred by the supply chain members and the
entire chain system under the decentralized and the semi-integrated channels?
(Q3.3) will the integration between the manufacturer and the retailer always be
profitable? If not, when would the manufacturer like to choose to forward

53

Chapter 3

integrate with the retailer?

(Q3.4) how will the market related parameters influence the channel members
equilibrium prices and profits?
To address these questions, game theory is used to analyze equilibrium prices for
both the leader-follower and the independent power structures in the decentralized and
the semi-integrated channels. This chapter also studies the effects of the power
structures, the channel structures and the market parameters on the equilibrium prices
and profits of individual channel members and the supply chain system.
The remainder of this chapter is organized as follows. The following section gives
the notations and formulates the profit models for the supplier, the manufacturer and
retailer. Section 3.3 and Section 3.4 illustrate two non-cooperative game models for
the decentralized channel the semi-integrated channel respectively and give solutions
to these models. Section 3.5 presents the price model for the integrated channel. In
Section 3.6, we discuss the effects of power structures, channel structures and market
parameters on the equilibrium prices and profits. We summarize the chapter in Section
3.7.
3.2 MODEL FORMULATION AND NOTATIONS
We consider the supply chain of one supplier, one manufacturer and one retailer of
a product with price sensitive demand. The supplier provides the manufacturer with
the sole raw material used to produce a single product sold to the retailer. Then the
retailer sells the product to customers. This simple monopoly structure allows us to

54

Chapter 3

focus on the competition and coordination between different echelons, without the
distraction of multiple products, multiple suppliers, manufacturers and retailers.
Similar assumptions can be seen from Jeuland and Shugan (1983), Alan and Medini
(1992), etc.. We use s, m, r to index the supplier, the manufacturer, the retailer,
respectively.
Demand is assumed to be a function of the retailers retail price (pr) paid by end
customers. If demand is price sensitive with constant price elasticity, we employ the
following non-linear demand function:

D pr aprb , a 0 b, ,0

(3.1)

where a is a market scale parameter, and b is the price elasticity of the demand, which
is always positive. This is because b 0 implies that D increases at a diminishing
rate as pr decreases. This demand function is fairly common in marketing literature
(Abad 1994; Lee 1993; Li et al. 1996; Kim and Lee 1998; Yue et al. 2006; Joseph and
Zhang 2009).
We further assume that all customer demand for the retailer will be satisfied. We
study a one period static model. With the deterministic market demand, it is mild to
assume that the manufacturer has the capacity to produce enough to satisfy the
retailers demand and the supplier could also provide enough material for the
manufacturer.
Given the level of demand, to determine the profits of the retailer, the
manufacturer and the supplier, we assume the supplier provides its raw material at a

55

Chapter 3

price of ps and the manufacturer sells its product at a wholesale price pm. Let mm and
mr denote the manufacturers profit margin and the retailers profit margin,
respectively. Further, we denote the suppliers procurement cost per unit raw material
as cs and the production cost per unit product as cm. s is assumed to be the usage
amount of unit raw material per unit product. This means that if the manufacturer
produces D unit product, he will purchase s D from the supplier.
Assuming that the retailer controls the values of the retail price pr, the
manufacturer controls the values of the wholesale price pm and the supplier controls
the value of the raw material price ps. Then the retailers profit function is given as:

r pr mr D pr ,

(3.2)

where mr pr pm .
The manufacturers profit function is:

m pm mm D pr ,

(3.3)

where mm pm ps s cm .
The suppliers profit function is:

s ps ps cs s D pr .

(3.4)

Using the profit functions identified above, we then determine the optimal pricing
decisions of the retailer, the manufacturer and the supplier under different channel
structures and power structures.
3.3 DECENTRALIZED CHANNEL
In this section, we consider the decentralized channel structure, in which the

56

Chapter 3

manufacturer uses independent supplier and retailer. We consider two power balance
scenarios under this channel structure, leader-follower and independent scenarios. For
the first scenario, the manufacturer takes the channel leadership, while the supplier
and the retailer are the followers. For the second one, the supplier, the manufacturer
and the retailer are of independent equal status and no one dominates over others. As
shown in Figure 3.1, we use Stackelberg game structure to model the first scenario
and Nash game structure for the second one.

ps

pm

ps

pr

(a) Manufacturer Stackerlberg

ps

pm

pm

pr

pr

ps
(b) Vertical Nash

Figure 3.1. Game structures for the decentralized channel

3.3.1 Manufacturer Stackelberg

We use Stackelberg game to model the leader-follower power balance scenario. In

fact, it is a sequential game, composed of two Stackelberg games. For convenience,
we call this game model as Manufacture Stackelberg (MS). The first Stackelberg
game is between the manufacturer and the supplier. In this game, the manufacturer
chooses its profit margin using the reaction function of the supplier. The supplier sets
its raw material price, conditional on the manufacturers profit margin. The second
57

Chapter 3

Stackelberg game is between the manufacturer and the retailer, in which the
manufacturer chooses its profit margin using the retailers reaction function and the
retailer determines its profit margin given the manufacturers profit margin. The game
structure can be referred from Figure 3.1(a).
Under the above assumption, the manufacturer takes the suppliers and retailers
reaction functions into consideration for its pricing decision. We first solve the second
Stackelberg game. The retailers reaction function can be derived from the first-order
condition of (3.2):

D pr
r
D pr pr pm
0.
pr
pr

(3.5)

From (3.5), the retailers reaction function can be derived:

pr pr pm .

(3.6)

The supplier determines its raw material price given the manufacturers profit
margin mm. Using pm mm ps s cm and (3.6), we have:

pr pr mm ps s cm .

(3.7)

Substituting (3.7) into the profit maximization condition for the supplier:

D pr pr
s

0.
pr , ps s D pr ps cs s
ps
pr
ps

(3.8)

Then we can derive the reaction function for the supplier:

ps ps pr p's pm .

(3.9)

Substituting the supplier and the retailers reaction functions (3.6) and (3.9) into
the manufacturers profit maximization condition:

58

Chapter 3

D pr pr
m
p
1 s s D pr pm ps s cm

0.
pm
pm
pr
pm

(3.10)

We can obtain the Stackelberg equilibrium of the two games as a solution for the
Manufacturer Stackelberg model. The equilibrium prices and profits for this game
structure can be referred from Table 3.1.
3.3.2 Vertical Nash

The second independent power balance scenario is formulated as a Nash game. In

this game, the supplier, the manufacturer and the retailer make pricing decisions
simultaneously and non-cooperatively. Again for convenience, we call this game
Vertical Nash (VN). In this game, the supplier chooses its raw material price
conditional on the manufacturers profit margin and the retailers profit margin to
maximize its profit. The manufacturer chooses its profit margin conditional on the
suppliers raw material price and retailers profit margin. The retailer sets its profit
margin so as to maximize its profit conditional on the suppliers raw material price
and the manufacturers profit margin. The game is played according to the rule seen
from Figure 3.1(b).
The Nash equilibrium for the Vertical Nash model can be represented as a solution
for our pricing problem. The first-order condition for this equilibrium involves the
retailer profit maximization condition (3.5) and the following two profit maximization
conditions:

D pr
m
D pr pm ps s cm
0;
pm
pr
59

(3.11)

Chapter 3

D pr
s
s D pr ps cs s 2
0.
ps
pr
Substituting

the

D pr

with

non-linear

(3.12)
demand

function

(3.1)

and

simultaneously solving (3.5), (3.11) and (3.12), we have the results for optimal prices
and profits shown in Table 3.2.
3.4 SEMI-INTEGRATED CHANNEL
In the semi-integrated channel, the manufacturer chooses to integrate with either
the retailer or the supplier first and then works with the supplier or the retailer
independently. In effect, the supply chain with this channel structure is a two-level
system where the manufacturer integrates with another echelon to be a single decision
maker.
3.4.1 MR-integration channel

In this section, we consider the channel structure that the manufacturer integrates
forward with the retailer in the three-level supply chain. We call this channel structure
as MR-integration channel. Also, two power balance scenarios are considered for the
MR-integration channel, leader-follower and independent. The first one is the two
integrated chain members (the manufacturer and the retailer) act as the leader, while
the independent member (supplier) acts as the follower. The second one is that the
two integrated members and the independent member are of equal status. We
formulate Stackelberg and Nash games for the two scenarios respectively. Since the
manufacturer and the retailer integrate together, we assume that there in no transfer

60

Chapter 3

price between them. Hence, there is no need to specify the manufacturers price in the
modeling process.
3.4.1.1 MR-Stackelberg
We first consider the leader-follower power balance scenario that the manufacturer
and the retailer integrate and act as the leader of the supply chain, while the supplier
acts as the follower. We formulate Stackelberg game between the integrated
manufacturer and retailer and the independent supplier. We call this game model as
MR-Stackelberg (MR-S). The manufacturer and the retailer agree to make their own
profit margin decision taking the suppliers reaction function into account. The
supplier conditions its raw material price on the profit margin given by the
manufacturer and the retailer. The game structure is shown as Figure 3.2(a).
The profit function for the manufacturer and the retailer is:

mr mmr D pr

(3.13)

where mmr pr ps s cm .
mmr is the profit margin for the integrated manufacturer and retailer. The suppliers
reaction function can be derived from the first-order condition of (3.4):

D pr
s
s D pr ps cs s 2
0.
ps
pr

(3.14)

Then we can obtain the suppliers reaction function:

ps ps pr .

(3.15)

Taking (3.15) into account, the manufacturer can obtain its optimal pricing
decision through the following first-order condition of (3.13):
61

Chapter 3

D pr

mr ps
1
s D pr pr ps s cm
0.
pr pr

(3.16)

Substituting D pr with demand function (3.1), we have the Stackelberg

equilibrium results for this game structure on prices and profits in Table 3.1.

ps

pm

(a) MR-Stackelberg

ps
s

pm
m

(b) MR-Nash

pm

pr

(c) SM-Stackelberg

pm
s

pr
r

(d) SM-Nash

Figure 3.2. Game structures for semi-integrated channels

3.4.1.2 MR-Nash
The independent power balance scenario here features that the integrated
manufacturer and retailer are of equal power with the supplier. We formulate Nash
game between them and call them as MR-Nash (MR-N). The supplier chooses its raw
material price conditional on the profit margin given by the manufacturer and the
retailer to maximize its profit. The manufacturer and the retailer integrate to choose its
profit margin conditional on the suppliers raw material price to maximize their total

62

Chapter 3

profit. The game structure is shown as Figure 3.2(b). The equilibrium conditions for
Nash game can be derived from the first order conditions of (3.4) and (3.13).

D pr
s
s D pr ps cs s 2
0;
ps
pr

(3.17)

D pr
mr
D pr pr ps s cm
0.
pr
pr

(3.18)

Simultaneously solving (3.17) and (3.18), we have the Nash equilibrium results
for prices and profits shown as Table 3.2.
3.4.2 SM-integration channel

In this subsection, we study the channel structure that the manufacturer achieves
backward integration with the supplier in the three-level supply chain. We call the
channel structure as SM-integration channel. We also formulate non-cooperative
game between the integrated manufacturer and supplier and the independent retailer.
Leader-follower and independent power balance scenarios are considered for the
SM-integration channel. The first one is that the integrated members act as the leader,
while the independent retailer acts as the follower. The second one is that the two
integrated members and the independent retailer are of equal status.
3.4.2.1 SM-Stackelberg
We consider the power balance scenario that the manufacturer and the supplier
integrate and act as the leader of the supply chain, while the retailer acts as the
follower. We formulate Stackelberg game between the integrated manufacturer and
supplier and the independent retailer. We mark the game as SM-Stackelberg (SM-S).
63

Chapter 3

In this game, the supplier and the manufacturer agree to determine their wholesale
price taking the retailers reaction function into account. The retailer conditions its
retail price on the wholesale price given by the supplier and the manufacturer. The
game structure is shown as Figure 3.2(c).
The profit function for the supplier and the manufacturer is:

sm pm cm cs s D pr .

(3.19)

Substituting the retailers reaction function (3.6) in to the first order condition of
(3.19), we have:

D pr pr
sm
D pr pm cm cs s

0.
pm
pr
pm

(3.20)

Solving (3.20), we have Stackelberg equilibrium results for prices and profits
shown in Table 3.1.
3.4.2.2 SM-Nash
This part illustrates the independent power balance scenario for the integrated
manufacturer and supplier and the independent retailer. We formulate Nash game
between them. This game is marked as SM-Nash (SM-N). In this game, the retailer
chooses its retail price conditional on the wholesale price given by the supplier and
the manufacturer to maximize its profit. The supplier and the manufacturer integrate
to choose the wholesale price conditional on the retail price to maximize their total
profit. The game structure can be seen from Figure 3.2(d). The equilibrium conditions
for Nash game can be derived from the first order conditions of (3.2) and (3.19).

64

Chapter 3

D pr
sm
D pr pm cm cs s
0.
pm
pr

(3.21)

Simultaneously solving (3.5) and (3.21), we have the Nash equilibrium results
shown as Table 3.2.
3.5 INTEGRATED CHANNEL
In this section, we focus on the integrated channel (marked as I). In this channel,
the supplier, the manufacturer and the retailer integrate together to take decisions to
maximize the entire system profit. The full vertical integration of the supply chain
wide prevents the manufacturer from dealing with the conflicting incentives that an
independent supplier or retailer would have. We assume there is no transfer price
between the supplier, the manufacturer and the retailer, and thus only a single retail
price pr is to be determined.
The profit function for the total three members is:

pr cm cs s D pr .

(3.22)

Taking the first order condition of (3.22):

D pr

D pr pr cm s cs
0.
pr
pr

(3.23)

Through (3.23), we can obtain the optimal retail price. The other results for the
integrated channel can be referred from Table 3.1.
All the related quantities are summarized in Table 3.1 and Table 3.2. Table 3.1
shows the results of the leader-follower power structure. Table 3.2 shows the results
for independent power structure. For the convenience of comparison, the results for

65

Chapter 3

the integrated channel are included in the two tables. Besides, in order to make the
results meaningful, we assume that price elasticity b is greater than 3 (see Table 3.2).
3.6 DISCUSSION
This section discusses several implications that are observed from the results. We
focus particularly on the effects of power structures, channel structures and market
parameters. In the following discussion, we use superscript MS, VN, MR-S, MR-N,
SM-S, SM-N and I to denote the corresponding quantities for the MS (Manufacture
Stackelberg), VN (Vertical Nash), MR-S (MR-Stackelberg), MR-N (MR-Nash), SM-S
(SM-Stackelberg), SM-N (SM-Nash) and I (integrated) cases, respectively.
3.6.1 Effects of power structure

Choi (1991; 1996) studies the effect of power structures on the equilibrium prices
and profits of the channel members in a traditional channel composed of the
manufacturer and the retailer and shows that under non-linear demand function, when
no one takes the channel leadership, each member will lose. Here, we will discuss the
effect of different power structures on the equilibrium prices and profits in the above
three-level supply chain. Integrated channel is not discussed since it does not involve
different power structures. The following Proposition illustrates the effects of the two
different power structures of the decentralized channel and MR-integration channel
respectively.
Proposition 3.1: a) For the decentralized channel, all the supply chain members and
the entire system prefer the MS case to the VN case for the lower equilibrium prices
66

Chapter 3

and the larger profits.

b) For the MR-integration channel, MR-S case is preferred by all the supply
chain members and the entire system.
Proof. We assume b 3 . Compare the retail price, the wholesale price and the raw
material price in the MS case with those in the VN case:
2
prMS pmMS b b 3
b3 3b2

1;
3
prVN
pmVN
b 1 b3 3b2 3b 1

(3.24)

psMS
s cs cm
1 b 1
0.
2
VN
ps
b 1 b 2 s cs cm

(3.25)

Hence, we have the relationships: prMS prVN , pmMS pmVN , psMS pVN
. The
s
equilibrium prices for all supply chain members in the MS case are not less than those
in the VN case.
Compare the retailers, the manufacturers, the suppliers profits and the entire
system profit in the MS case with those in VN case:

MS b 1
Obviously, rVN 2

b b 3
r

b 1

b3 3b 2 3b 1

b3 3b 2

b 1

1.

(3.26)

b 1
b 1
mMS
,
this
is
because
is decrement function of b

1
b 1
mVN b2b b 3b 1
b 2b b 3
3b 1

3b 1

(see Appendix A). When b tends to infinite,

mMS
tends to the lowest value 1. That is,
mVN

b 1
lim 2b
1.
b 1
b
b b 3
3b 1

b 1

sMS
MS
1 . Thus, VN 1 .
Similarly, we have: VN 2b 1
b 1

s
b b 3
3b 2

So, the profits for all chain members and the entire supply chain system in the MS
67

Chapter 3

case are no less than those in the VN case: rMS rVN , mMS mVN , sMS sVN ,

MS VN . This completes the proof of part (a). We can refer from Appendix B for
the proof of part (b).
From Proposition 3.1, we can see that, when non-linear demand function (3.1) is
employed, for the decentralized or the MR-integrated system, the leader-follower
power scenario is preferred by the supply chain compared with the independent power
scenario. That is, the manufacturer or the integrated manufacturer and retailer would
rather take the leadership of the decentralized channel or the MR-integration channel.
Proposition 3.1 is also consistent with the results for the traditional channel structure
(Choi 1991). For SM-S and SM-N cases, the results are similar with those of MR-S
and MR-N cases. Discussions are omitted here.
3.6.2 Effects of channel structure

With channels of distribution changing rapidly, the importance of channel

selection has been emphasized. Different channel structures will influence the pricing
decisions and profits for all the supply chain members. Supply chain members prefer
to select the channel structure that could bring larger profits to them. In this
subsection, we study the effects of different channel structures of the three-level
supply chain. We first propose the following proposition.

Proposition 3.2: Compared with the decentralized channel, the manufacturers

forward integration with the retailer can always provide larger profits for all the

68

Chapter 3

supply chain members and the entire system when price elasticity b satisfies
2b2 2b 1 (b 1)3
2

b2
b (b 2)

b 1

1 or b 3.5396 .

Proof. Proposition 3.1 shows that in the decentralized channel, the MS case provides
larger profits for all the chain members and the entire system than the VN case and
the MR-S case provides larger profits than the MR-N case in the MR-integration
channel. Thus, we just need to compare the profits of the individual supply chain
members and the entire system in the MR-N case with those in the MS case. If the
MR-N case could provide larger profits than the MS case, the MR-S (or MR-N) case
will also have larger profits than the MS (or VN) case. That is the integration of the
manufacturer and the retailer can always provide larger profits for all the chain
members and the entire system.
Compare the joint profit of the retailer and the manufacturer in the MS case with
that in MR-N case:

rMS mMS b 1 2b 2b 1
.

b 1
MR N
mr
b 2b b 2
3b 3

b 1

3b 3

2b

2b 1

b 2b b 2

b 1

(3.27)

is a decrement function of b. This proof is similar to that

in Appendix A. Some numerical methods, such as bisection method, Newtons method,

can be employed to find the root of

b 1

3b 3

2b

2b 1

b 2b b 2

bisection method and find the root, b 3.5396 .

Therefore,

69

b 1

1 . Here, we use

Chapter 3

MS
r

MS
m

MR N
mr

, when

b 1

3b 3

2b

2b 1

b 2b b 2

b 1

1 or b 3.5396 . (3.28)

Compare the suppliers profit in the MS case with that in MR-N case:

sMS
(b 1)3b 2

sMR N b2b 1 (b 2)b 1

(3.29)

sMS
It is a decrement function of b. When b tends to 3, MR N tends to the highest
s
value 0.5267. Hence, we have sMS sMR N .
Similarly, the entire system of the MS case and that of the MR-N case have the
following relationship: MS MR N . This completes the proof of Proposition 3.2.
McGuire and Staelin (1983) show that vertically integration can maximize joint
profits in a monopoly for the channel composed of the manufacturer and the retailer.
From Proposition 3.2, we can see that the integration for the manufacturer and the
retailer can not always maximize the joints profit even in a monopoly. The main
reason for the difference between McGuire and Staelins results and the result from
this research is that we consider the integration in a three-level channel, which is
different from the traditional channel in McGuire and Staelins study (1983). In the
MR-integration channel, the retailer charges no more retail price, while the supplier
charges no less raw material price compared with the prices in the MS case. For
example, prMR N prMS , psMS psMR N . Because:
prMS
b 2 (b 2)
b3 2b 2

.
3 2 5
prMR N
(b 1)3
3
2
b 2b (b )
2
4

3
5
Since we assume b 3 , thus, (b ) 2 0 . Hence, prMR N prMS .
2
4
70

(3.30)

Chapter 3

psMS
(b 2)(b2 b 1) s cs b(b 2)cm
s cs cm
1
1
0.
MR N
3
2
3
ps
(b 1) s cs (b 1) cm
(b 1) s cs (b 1) 2 cm
(3.31)
So, psMS psMR N .
Thus, for the MR-integration channel to lead to improved profits, the market
demand must be sensitive enough to the retail price. That is, price elasticity b must be
adequately large, so that the reduction in retail price could enhance market demand
and bring additional profits adequate to cover the loss caused by the higher raw
material price. Therefore, the integration of the manufacturer and the retailer could
result in a larger joint profit to them.
Lau and Lau (2003) define the channel efficiency (marked as CE) with the ratio of
the channel profits to the integrated channel profits. By this definition, the channel
efficiency for all the cases in this chapter is shown as the last rows of Table 3.1 and
Table 3.2. From the decentralize channel, the semi-integrated channel to the integrated
channel, we consider the supply chain integration to be incremental. The following
proposition is proposed:
Proposition 3.3: An increased supply chain integration leads to higher channel
efficiency.
Proof. From Proposition 3.1, we can see that MS VN and MR S MR N . Hence,
2
CE MS 3b 3b 1 b 1

b 1
CEVN
3b2b 2 b 3

3b 3

2
CE MR S 2b 1 b 1

CE MR N
2b b b 2

MS
1;
VN

MR S
1.
MR N

b 1

71

Chapter 3

From Proposition 3.1, we can see that MS MR N . So,

2b 1 bb

MR N
CE MR N

1.
CE MS
3b2 3b 1 b 1b1 MS

Then, we just need to compare the channel efficiency of the integrated channel
with that of the MR-S case:

CE I
bb

CE MR S 2b 1 b 1b1
bb

2b 1 b 1

b 1

is decrement function of b. When b tends to 3,

CE I
tends
CE MR S

to the lowest value 1.35.

So, CEVN CE MS CE MR N CE SM N CE MRS CE SM S CE I 1 .
This completes the proof.
Thus, the integrated channel has the highest channel efficiency, followed by the
semi-integrated channel, and then the decentralized channel. Considering Proposition
3.2, the semi-integrated channel may not necessarily increase the profits of individual
chain members. However, from the supply chain system wide, this integration brings
larger profit for the supply chain system comparing the decentralized channel.
3.6.3 Effects of market related parameters

The market environment has great influence on the firm strategies and the
performance of the entire supply chain (Park and Mason 1990; Venkatraman and
Prescott 1990). The effect of market parameters have been studied by many empirical
works (Prafulla 2006; Joseph and Zhang 2009). They show that market parameters as

72

Chapter 3

a major factor influence the pricing and profits of the supply chain members. In this
subsection, we consider the effects of the scale parameter a and price elasticity b on
the pricing decisions and profits of the supply chain members under the decentralized
channel structure. The semi-integrated and integrated channels are not analyzed here
since their price and profit structures are much similar with those of the decentralized
channel. Table 3.3 and Table 3.4 summarize the effects of market scale parameter a
and price elasticity b on the equilibrium prices and profits for the MS and VN cases.
A larger market scale implies a better market environment, while the degree of
benefits for individual chain members depends on the underlying power structure. The
following propositions summarize the major findings of the prices and profits as the
change of market scale under the decentralized channel.
Proposition 3.4: a) No channel members would change their equilibrium prices as
market scale parameter a changes.
b) As market scale parameter a increases, the retailer benefits most and the
manufacturer benefits the least in the MS case; in the VN case, each member has the
same profit increase
c) Market scale parameter a has a more significant effect on the MS case than that
on the VN case.
Proof. In our decentralized channel, we consider the first order partial derivatives of
price and profit with respect to a. For the MS case and VN case, as Table 3.3 shows,
we have:

73

Chapter 3

pr MS pm MS ps MS
prVN pmVN psVN

0,

0.
a
a
a
a
a
a

This completes the proof of part (a).

mVN sVN rVN
r MS s MS m MS

0,

0.
a
a
a
a
a
a

This completes the proof of part (b).

rMS rVN mMS mVN sMS sVN MS VN

,
,
,
.

a
a
a
a
a
a
a
a
This completes the proof of part (c).
Proposition 3.4(a) shows that the change of market scale would not affect the
chain members pricing decisions. Hence, the increased market scale will obviously
increase the chain members profits for the larger market demand. From Proposition
3.4(b), we can see that when the manufacturer takes the channel leadership, his profit
increases the least and the retailers profit increases the most as the market scale
becomes larger. When independent power structure is employed, the change of market
scale has the same effects on the chain members. Proposition 3.4(c) indicates that
when the manufacturer takes the channel leadership, the chain members profits are
more sensitive to the change of the market scale.
Price sensitivity is fundamental to many important aspects of retail policy, such as
pricing, promotions (Hoch et al. 1995). We then consider the influence of price
elasticity b on the equilibrium prices and profits for the MS and VN cases.
Firstly, take the first order partial derivatives of profits with respect to b, for the
MS case, we have:

74

Chapter 3

r MS
a (b 1)3
3b 2

b
b
s cs cm

b 1

r MS

m MS a(b 1)2b b 1

b
b3b
s cs cm
s MS a(b 1)b (b 1)2

b
b3b 1 s cs cm

b 1

m MS ,

(3.32)

b 1

s MS

1 2
where r MS 3ln 1 b 1 s cs cm ,
b b

1
m MS 3ln 1
b 1 s cs cm ,
b b 1

1 2b 1
s MS 3ln 1
b 1 s cs cm .
b b(b 1)

For the VN case:

rVN mVN sVN a b 3

b
b
b
b s cs cm

b 1

VN ,

(3.33)

3
2

b 1 s cs cm .
where VN ln 1
b b3

Proposition 3.5: a) An increased price elasticity b results in the reduced equilibrium

price for each supply chain member; the reduction of the retail price is larger than
that of the wholesale price.
b) In the MS case, if b satisfies r MS 0 the profit for each supply chain member
increases as price elasticity b increases; if s MS 0 , no one could benefit from an
increased b. In the VN case, if VN 0 , the increase of price elasticity b will benefit
all the supply chain members; otherwise, all may lose.
Proof. Referred from Table 3.4, the first order partial derivatives of prices with

75

Chapter 3

respect to b have the following relationship:

ps MS
pr MS pm MS
0.

0,
b
b
b

(3.34)

psVN
prVN pmVN
,
0.

0
b
b
b

(3.35)

For the MS case and VN case, the equilibrium prices are inversely proportional to
the change of parameter b. Form (3.34) and (3.35), we can also deduce
p MS
pr MS
m
b
b

or

p VN
prVN
m . Hence, we conclude that the change in retail price
b
b

is larger than the change in the wholesale price. This completes the proof of part (a).
Since from (3.32), we have that:

s MS m MS r MS .

(3.36)

We know right away that r MS 0 implies s MS 0 and m MS 0 . Thus,

m MS
s MS
r MS
0;
0,
0 . s MS 0 implies m MS 0 and r MS 0 . So,
b
b
b
m MS
MS
r MS
0;
0, s 0 .
b
b
b
For the VN case, VN 0 implies

rVN mVN sVN

0 .
b
b
b

This proposition suggests that when the market demand becomes more sensitive
to the retail price, all the supply chain members will reduce their equilibrium prices
and the retailer has a more price reduction than the manufacturer. Meanwhile, the
change of the profits of the individual chain members depends on the value of price
elasticity b. Proposition 3.5(b) presents the distinguish conditions for the variation

76

Chapter 3

tendency of the chain members profits with the changing of price elasticity b.
The following corollary can be obtained from Proposition 3.5:
Corollary 3.1: When product cost (raw material cost and production cost)

s cs cm 0.2749 , each chain members profit will increase as price elasticity b

increases in the MS case.

1 2
Proof. Notice that r MS 3ln 1 b 1 s cs cm .
b b

1 2
3ln 1 b b

MS
If s cs cm max
satisfies, then r 0 .
b

1
1 2
We assume that b is greater than 3. Since 3ln 1 and
are
b 1
b b

1 2
3ln 1
b b is
decrement functions and both of them are larger than 0,
b 1

decrement function of b.

1 2
1 2
3ln 1
3ln 1

b b lim
b b 0.2749 . From (36),
Hence, max
b 3

b 1
b 1

all the chain members profits will increase. This completes the proof of Corollary 3.1.

Corollary 3.1 illustrates the relationship between the product cost and the profit
change when price elasticity b increases in the MS case. That is, when the
manufacturer takes the channel leadership, if the product cost is no less than 0.2749,

77

Chapter 3

the chain members profits will increase as the market becomes more sensitive to the
retail price.
3.7 SUMMARY
Most previous studies have focused on price coordination in traditional channel
structures with two echelons (manufacturer and retailer, or manufacturer and supplier).
In the contemporary real market, coordination for the multi-level supply chain with
different power structures and channel structures is inevitable and necessary. In this
chapter, we extend the growing literature of channel studies by analyzing pricing
strategies for a three-level supply chain consisting of supplier, manufacturer and
retailer. Three different channel structures are considered. They are the decentralized
channel, the semi-integrated channel and the integrated channel. Two non-cooperative
games are used to model different power structures for the first two channel structures,
i.e., Stackelberg and Nash games. We also investigate the effects of power structures,
channel structures and market parameters on the pricing decisions and profits for
channel members.
Our studies show that when the manufacturer or the integrated members take the
leadership of the supply chain for either decentralized or semi-integrated channel, the
equilibrium prices are lower and profits are higher compared with the channel without
such leadership.
In this chapter, we also provide fresh new insights into vertical integration.
McGuire and Staelin (1983) conclude that vertical integration can maximize joint

78

Chapter 3

profits in a monopoly. In MR-integration channel, the integration of the manufacturer

and the retailer cannot provide larger profits for them or the other channel member
unless the price elastic b is no less than a certain level (3.5396). In general, the
integration for the manufacturer and the retailer cannot always improve their profits in
a monopoly when a multi-level channel is considered.
Another major finding of this chapter is that when the manufacturer takes the
channel leadership, the chain members profits will increase as the market becomes
more sensitive to the retail price, if the product cost is greater than a certain level
(0.2749).

79

Table 3.1. Results for leader-follower power structure

Decentralized
Semi- integrated
Manufacturer Stackelberg
MR Stackleberg (MR-S) SM Stackleberg (SM-S)
(MS)
b3 ( s cs cm )
b2 ( s cs cm )
b2 ( s cs cm )
pr
(b 1)3
(b 1)2
(b 1)2
pm

b2 ( s cs cm )
(b 1)2

ps

(b2 b 1) s cs bcm
s (b 1)2

(b2 b 1) s cs bcm
s (b 1)2

b 1

80

a (b 1)2 b b 1

b3b
s cs cm

a (b 1)b (b 1)2

b3b1 s cs cm

a (b 1) (b 1)2

b2b s cs cm

b 1

b 1

a (3b2 3b 1) (b 1)3

b 3b
s cs cm

(3b2 3b 1)(b 1) 2( b1)

b2 b

a (b 1)2

b2b1 s cs cm
b 1

b 1

b 1

a (2b 1) (b 1)2

b2 b s cs cm

(2b 1)(b 1)b1

bb

b 1

a (b 1)2

b2b1 s cs cm

b 1

a (b 1) (b 1)2

b2b s cs cm

b 1

a (2b 1) (b 1)2

b2 b s cs cm

(2b 1)(b 1)b1

bb

b 1

a b 1

bb s cs cm

b 1

a b 1

bb s cs cm

b 1

1
Chapter 3

(System
Profit)
System
efficiency
(CE)

b( s cs cm )
b 1

b( s cs cm )
b 1

a (b 1)3

b3b2 s cs cm

Integrated (I)

Table 3.2. Results for independent power structure *

Decentralized
Semi- integrated
Vertical Nash (VN)
MR Nash (MR-N)
SM Nash (SM-N)
b( s cs cm )
b3

pm

(b 1)( s cs cm )
b3

ps

(b 2) s cs cm
s (b 3)
a b3

bb s cs cm

b 1

b 1

a b3

bb s cs cm

b 1

a b3

bb s cs cm

a b2

bb s cs cm

3a b 3

bb s cs cm

b 1

2a b 2

bb s cs cm

81

pr

(System
Profit)
System
efficiency
(CE)

b 3
3

b 1

b 1

b( s cs cm )
b2

b( s cs cm )
b2

Integrated (I)
b( s cs cm )
b 1

(b 1)( s cs cm )
b2
(b 1) s cs cm
s (b 2)
a b2

bb s cs cm

b2
2

b 1

b 1

b 1

b 1

b 1

a b2

bb s cs cm
a b2

bb s cs cm

b 1

b 1

2a b 2

bb s cs cm

b2
2

b 1

b 1

a b 1

bb s cs cm

b 1

a b 1

bb s cs cm

b 1

b 1

1
Chapter 3

* In order to make the results for Vertical Nash (VN) case meaningful here, we assume the price elastic b is greater than 3.

Table 3.3. Effects of market scale parameter a

Decentralized
Manufacturer Stackelberg (MS)
Vertical Nash (VN)
pr
a

pm
a

ps
a

0
b 1

82

1 b3

bb s cs cm

b 1

b 1

1 b3

bb s cs cm

b 1

b 1

1 b3

bb s cs cm

b 1

r
a

1 (b 1)3

b3b 2 s cs cm

m
a

(b 1)2b b 1

b3b s cs cm

s
a

(b 1)b (b 1)2

b3b 1 s cs cm

Chapter 3

Table 3.4. Effects of price elasticity b

Decentralized
Manufacturer Stackelberg (MS)
pr
b

pm
b

83

r
b
m
b
s
b

3b 2 ( s cs cm )
(b 1) 4

ps
b

a (b 1)3

b3b 2 s cs cm

b 1

a(b 1)2b b 1

b3b
s cs cm
a(b 1)b (b 1) 2

b3b 1 s cs cm

2b( s cs cm )
(b 1)3

(b 1)( s cs cm )
s (b 1)3

3( s cs cm )
(b 3)2

2( s cs cm )
(b 3) 2

a b3

bb s cs cm

b 1

1
3ln 1 b b 1 b 1 s cs cm

a b3

bb s cs cm

b 1

1 2b 1
b 1 s cs cm
3ln 1
b b(b 1)

a b3

bb s cs cm

b 1

1 2
3ln 1 b b b 1 s cs cm

b 1

b 1

Vertical Nash (VN)

s cs cm
s (b 3)2

3
2

ln 1 b b 3 b 1 s cs cm

3
2

ln 1 b b 3 b 1 s cs cm

3
2

ln 1 b b 3 b 1 s cs cm

Chapter 3

CHAPTER 4
GAME-THEORETIC COORDINATION OF PRICING AND INVENTORY
DECISIONS IN A SUPPLY CHAIN WITH RETAILER COMPETITION

4.1 INTRODUCTION

In the previous, coordinating pricing decision in multi-level supply chain has been
studied. For many industries, the coordination of price decisions with inventory leads
to the improvement of the efficiency of both the supply chain and individual firms
(Weng 1995; Chan et al. 2004). This chapter considers coordinating pricing and
inventory decisions of supply chain problem (CPISC) in a two-echelon supply chain
composed of one single manufacturer and multiple retailers.
Coordinating pricing and inventory decisions has been studied by many
researchers. Weng and Wong (1993) and Weng (1995) propose a model of
supplier-retailer relationship and confirm that the coordination of pricing and
inventory decisions benefits both supply chain system and individual firms. Prafulla et
al. (2006) present a set of models of coordination for pricing and order quantity
decisions in a one manufacturer and one retailer supply chain. They also discuss the
advantages and disadvantages of various coordination possibilities. Different from
these literature focused on one wholesaler and one retailer situations, Boyaci and
Gallego (2002) study pricing and inventory policies that maximize the channel profits
in the supply chain with a wholesaler and one or many retailers. None of these studies

84

Chapter 4

takes the interaction and competition between different retailers into account.
However, retailer competition influences retailers pricing of brands significantly
(Chintagunta 2002).
Esmaeili et al. (2009) proposed several game models of seller-buyer relationship
to optimize pricing and lot sizing decisions. Yu et al. (2009) simultaneously consider
pricing and order intervals as decision variables using Stackelberg game in a supply
chain with one manufacturer and multiple retailers. Although game theory approach is
employed to coordinate pricing and inventory policies in the above research, the
authors focus on the relationship with seller and buyer of different market power, for
example the manufacturer dominates the retailers. Little attention has been given to
the market structure in which the retailers retain equal as the manufacturers (Buzzell
et al. 1990). In most cases, the concept of Nash equilibrium is more general and useful
to understand the behavior of the individual supply chain members.
In this chapter, we study the CPISC problem in a two-level supply chain with one
single manufacturer and multiple competing retailers, where the manufacturer
purchases different types of raw materials used to produce substitutable products for
different retailers in the same market. In this supply chain, all the supply chain
members are rational, of equal market power and determine their pricing and
inventory decisions non-cooperatively.
The objective in this chapter is mainly to determine the optimal pricing and
inventory decisions for the manufacturer and the retailers to maximize their individual

85

Chapter 4

profits. We raise the following research questions. The questions faced by the retailers
include:
(Q4.1) what are the optimal retail prices for products sold to different markets?
(Q4.2) how frequent should the product procurement take place?
For the manufacturer, he has to decide on:
(Q4.3) what are the optimal product wholesale prices for different retailers?
(Q4.4) how frequent should the production process take place?
(Q4.5) how frequent should the raw material procurement take place?
We develop a dynamic game-theoretic framework for this coordination problem.
In this game framework, the coordination problem is modeled as a dynamic
non-cooperative game with complete and perfect information, where the competing
retailers formulate a subgame and play the whole game with the manufacturer. This
game settles non-cooperative equilibrium (Nash equilibrium) solution such that any
chain member cannot improve his profit by acting unilaterally without degrading the
performance of the other players. Analytical and computational method to obtain
equilibrium of this game is also devised. Finally, we present a numerical illustration to
justify the proposed game model and the proposed solution algorithm. Several
sensitivity analyses are conducted to examine the influence of the market and
production related parameters on the decisions and profits of all the chain members.
The remainder of this chapter is organized as follows. We give the assumptions
and notations underlying our models in Section 4.2. This section also formulates the

86

Chapter 4

problem. In Section 4.3, the dynamic non-cooperative game model is proposed.

Section 4.4 develops the analytical and computational methods to solve the problem.
Then, in section 4.5, a numerical example and corresponding sensitivity analysis for
some selected parameters have been presented. Finally, this chapter concludes in
Section 4.6.

4.2 ASSUMPTIONS AND NOTATIONS

4.2.1 Problem description and assumptions

We consider a seller-buyer supply chain, in which the manufacturer purchases raw
materials outside to produce substitutable products and delivers them to the different
retailers. The substitutability of different products results in the competition between
the retailers. The non-cooperative retailers reach an agreement and as a whole
negotiate with the manufacturer on pricing and inventory decisions to maximize their
own profits. When the equilibrium is achieved between them, the manufacturer will
setup his production and the retailers will then purchase these products and distribute
to their customers. We give the following assumptions:
(A4.1): Each retailer only sells one type of product. This production-sale inventory
model has been implicitly made by Lu (1995). The annual demand function for each
retailer is a decreasing and convex function with respect to its own retail price.
(A4.2): The replenishment time of raw materials is assumed to be integral multipliers
of the manufacturer setup time interval which is also the integer multipliers of all the

87

Chapter 4

retailers replenishment time. This integer multipliers mechanism has also been
adopted in many previous works (Goyal 1977; Moutaz 2002), which are more general
than the inventory policy in other works, such as lot-for-lot procurement policy
(Sarker and Parija 1994), VMI policy (Woo et al. 2001).
(A4.3): The raw materials used for different products are different.
(A4.4): Shortage are not allowed for the manufacturer, hence the annual production
capacity is greater than or equal to the total annual market demand (Esmaeili et al.
2009).
(A4.5): The manufacturer and the retailers are rational decision makers and have
equal market power.
4.2.2 Notations
All the input parameters and variables used in our models will be stated as follows.
Assume the following relevant parameters for the retailer:
L: Total number of retailers

rl : Index of retailer l

Rrl : Retailer ls annual fixed costs for the facilities and organization to carry this
product

Arl : A constant in the demand function of retailer l, which represents his market
scale

erll : Coefficient of the products price elasticity of the demand for retailer l

erlj : Coefficient related to the degree that product j could be substitutable by

88

Chapter 4

product l

hrl : Holding costs per unit of product l for retailer l

prl : Retail price charged to the customers by retailer l
Drl : Retailer ls annual demand
The retailers decision variables are:

Grl : Retailer ls profit margin

krl : The integer divisor used to determine the replenishment cycle of retailer l
The manufacturers relevant parameters are:
m: Index of the manufacturer
Vl: Total number of raw materials used for product l

slvl : Index of raw material vl used for product l, vl =1,2,,Vl

hmpl : Holding costs per unit of product l
hmslv : Holding costs per unit of raw material vl used for product l
l

s : Usage of raw material vl for a unit product l

lvl

S m : Setup cost per production

Oml : Ordering processing cost of raw materials used for product l per order
Pl : Annual production capacity product l, which is a known constant
Rm : Manufacturers annual fixed costs for the facilities and organization for the
production of this product

pml : Wholesale price charged by the manufacturer to the retailer l

cml : Production cost per unit product l

89

Chapter 4

cslv : The purchasing cost of raw material vl used for product l

l

The manufacturers decision variables are:

Gml : Manufacturers profit margin for product l

T : Manufacturers setup time interval

nml : The integer multiplier used to determine the procurement cycle of raw
materials used for product l

4.3 MODEL FORMULATION

4.3.1 Game scheme

In our game framework, the competing retailers are modeled as a subgame (RS
game) and formulates the whole MR game with the manufacturer. According to Kim
et al. (2003), the proposed MR game is, therefore, a dynamic game where the
manufacturers strategies in MR game affect the retailers strategies from RS game
and, at the same time, the decisions from RS game also affect those decisions in the
MR game. Through the dynamic interactions between the MR game and the RS game,
individual manufacturer and retailers could determine optimal pricing and inventory
decisions to maximize their payoffs. However, we can see this game is partially
dynamic, because it is not a sequential game between the manufacturer and the
retailers. They move simultaneously (i.e., statically) in the game.
The MR game in this chapter is also a non-cooperative game with complete
information where each player knows the other players strategy set but they are not

90

Chapter 4

involved in any binding agreements. That is they are aware that their decisions affect
each other and they take this into account when they make pricing and inventory
decisions. However, any agreement or commitment among them is not assumed to
take place. Since the RS game is nested in the whole MR game and both games are
also non-cooperative simultaneous games (i.e., Nash game), in this thesis, we can
refer the MR game to be a two-level nested Nash game. Figure 4.1 shows the
structure of the MR game.
Main game
MR game

Subgame
RS game
Figure 4.1. MR game structure

In order to present our game model, we define the following notations for the
whole MR game:

N m, r1 , r2 ,...rL is the set of players, i.e., one manufacturer and L retailers.

X m is the strategy set of the manufacturer. X rl is the strategy set of retailer l.

X r X r1 X r2

X rL is the strategy profile set of the retailers strategy sets and

the strategy space of the RS game. X X m X r is the strategy profile set of the
manufacturer and the retailers and the strategy space of the MR game.

xm X m denotes a strategy of the manufacturer, which consists of profit margin

Gml , setup time interval T and procurement decision for raw materials nml
91

Chapter 4

(l=1,2,L). xrl X rl is a strategy of retailer l, composed of profit margin Grl and

replenishment decisions krl .

xr xr1 , xr2 ,

, xrL X r is a strategy profile and consists of the strategies of the

retailers. x xm , xr X is a strategy profile, composed of the strategies of the

manufacturer and the retailers.

Z m and Z rl are the payoff functions of the manufacturer and retailer l.

4.3.2 The retailers model
We first consider the payoff function Z rl for the retailer l. The retailers objective
is to maximize his net profit by optimizing his strategy xrl , including profit margin

Grl and replenishment decision krl .

Linear demand functions have been used extensively in the marketing and
economics literatures for theoretical and empirical analyses of differentiated
oligopolies and monopolistically competitive industries (e.g., Samuelson 1947,
Pashigian 1961, Staelin and Winer 1976, McGuire and Staelin 1983 and Jeuland and
Shugan 1988, Choi 1996). Considering the demand functions in McGuire and Staelin
(1983) and Choi (1996), we define the following demand function:
D( prl ) Arl erll prl

j 1,2,..., L
j l

erlj prj , with Arl 0 , erll 0 , erlj 0 ; j, l=1,2,L. (4.1)

where Arl is a market scale constant, erl is coefficient of the products price
elasticity, and erlj is coefficient related to the substitution degree between products.
Because the retailers products are substitutable, we have erlj 0 (Samuelson 1947).
It can be seen that this demand function is downward sloping and a convex function
92

Chapter 4

with respect to prl , since

D( prl )
prl

0 , l=1,2,L.

As indicated in the second point of the assumption in section 4.2.1, the integer
multipliers mechanism is employed between the manufacturer and the retailers. Since
the setup time interval for the manufacturer is assumed to be T, the replenishment
cycle for retailer l is

cost is

TDrl
2krl

T
. krl should be a positive integer. Thus, the annual holding
k rl

hrl (see Figure 4.2(a)) and the ordering process cost is

Orl krl
T

. The

retailer l faces the holding cost, the ordering cost and an annual fixed cost. Therefore,
the retailer ls objective function is given by:

max Z r Gr Dr
rl , krl

TDrl
2krl

hrl

Orl krl
T

Rrl ,

(4.2)

Subject to

krl {1, 2,3,...} ,

(4.3)

Grl prl pml ,

(4.4)

D( prl ) Arl erll prl

j 1,2,..., L
j l

erlj prj ,

(4.5)

Grl 0 ,

(4.6)

0 Dr l Pl .

(4.7)

Constraint (4.3) gives the value of the divisor used to determine the retailer ls
replenishment cycle time. Constraint (4.4) indicates the relationship between the
prices (the retail price and the wholesale price) and retailer ls profit margin.
Constraint (4.6) ensures that the value of Grl is nonnegative. Constraint (4.7) gives
93

Chapter 4

the bounds of the annual demand, which cannot exceed the annual production
capacity of the product.
4.3.3 The manufacturers model
The manufacturers objective is to determine his strategy xm , composed of the
profit margins for all the products Gml , the setup time interval for production T and
procurement decision for raw materials nml , to maximize his payoff Z m .
The manufacturer faces annual holding costs, setup and ordering costs, and an
annual fixed cost. The annual holding cost for the manufacturer is composed of two
parts: the cost of holding raw materials used to convert to the products, the cost of
holding products. The manufacturer sets up the production once in each time interval
T. For any product l, the production run period is

TDrl
Pl

. The raw materials will be

consumed continuously during each production period, and the left will be kept until
the next production run starts (as Figure 4.2(b)). Therefore, the manufacturers raw
materials stock used for product l per procurement cycle consists of nml triangles and

ml

1 rectangles. According to Woo et al. (2001), the annual inventory cost of raw

material vl per unit time is

2P

hmslv

s Dr
lvl

ml

1 hmslv
2

s Dr T . The inventory
lvl

level of product l is shown as Figure 4.2(c). The annual inventory cost of product ls is
given by

1 Dr
Drl 1 l (as suggested by Lu (1995)). The setup cost S m
kr
2
Pl
l

hmpl T

occurs at the beginning of each production period. The manufacturer has to order
different raw materials for the production of each product, hence the ordering cost for
94

Chapter 4

the raw materials for product l is

Oml
nml T

. Thus, we can easily derive the

manufacturers payoff function m :

max Z m Gm Dr
l

Gml ,T ,nml
l 1,..., L

T
hmp Dr
2 l l l

hms
lvl slv Dl
l
l
vl 2 Pl

1 Dr
1 l
kr
Pl
l

n
T

ml

1 hmslv
2

S
Om
slvl DlT m l Rm
T
l nml T

(4.8)
Subject to

nml {1, 2,3,...} , for each l=1,2,L

Gml pml slv cslv cml , for each l=1,2,L
vl

(4.9)
(4.10)

Gml 0 , for each l=1,2,L

(4.11)

T 0.

(4.12)

Constraint (4.9) gives the value of the multiplier used to determine the
manufacturers procurement cycle time for raw materials need to produce product l.
Constraint (4.10) gives the relationship between the price (the wholesale price and the
raw material price) and the manufacturers profit margin. Constraint (4.11) and (4.12)
ensure that the values of Gml and T are nonnegative.

95

Chapter 4

Inventory level

TDrl / krl

T / krl

Time

(a) Retailer ls inventory level of final product

Inventory level

slv Drl T / Pl
nrl slv Drl T

nlT

Time

Inventory level
(b) Manufacturers inventory level for product l

TDrl / Pl

T / krl

(c) Manufacturers inventory level for product l

Figure 4.2. Inventory levels of retailer and manufacturer

96

Time

Chapter 4

4.3.4 MR game model

Based on the game structure shown in Figure 4.1 and the above payoff functions
of all the players, we can formulate the MR game model as follows:

M ax Z m

Gml ,T ,nml

Subject to
Constraints (4.9)-(4.12)

max Z r , for each l=1,2,,L

Grl , krl

Subject to
Constraints (4.3)-(4.7)
In this case, the retailers have the same cost structures, so we can use one uniform
formula to represent the payoff function Z rl of each retailer as defined in (4.2)-(4.7).
In the above game model, payoff function Z rl and constraints (4.3)-(4.7) for each
retailer l=1,2,,L in this game model formulate the RS game. Payoff function Z m
and constraints (4.9)-(4.12) formulate the MR game, which is the model of the
manufacturer.

4.4 SOLUTION ALGORITHM

Nash equilibrium is the most popular non-cooperative solution concept in game

theory, which is widely used for dynamic non-cooperative game (Basar and Olsder
1982). In this chapter, we also apply this Nash equilibrium concept to analyze the
strategies of the supplier, the manufacturer and the retailers in the whole MR game.
To obtain the Nash equilibrium, each player of the game takes the other players
decisions as given input parameters to determine his own decisions and adjusts his

97

Chapter 4

decisions corresponding to the changing of the other players decisions in order to

maximize his net profit. The process continues until no player is willing to change his
decisions, because any unilaterally changing will bring loss to him and then the Nash
equilibrium reaches.
Many researchers proposed solution algorithms to computing the Nash
equilibrium. For example, Liu (1998) proposes a genetic algorithm for finding the
Nash equilibrium. Periaux et al. (2001) merge genetic algorithms and Nash strategy to
make the genetic algorithm build the Nash equilibrium.
In this chapter, we mainly based on analytical theory used by Liu (1998) to
compute Nash equilibrium. In order to determine the equilibrium of the MR game, we
first use analytic method to calculate the best reaction functions of each player and
employ algorithm procedure to build the Nash equilibrium.
4.4.1 Reaction functions
4.4.1.1 The retailers reactions
We first solve the reaction functions of the retailers. We express the retailer ls
demand function by the corresponding profit margins. Substituting (4.4), (4.10), we
can rewrite (4.5) as:

Dr l Arl erll Grl Gml slv cslv cml erlj Grj Gm j s jv cs jv cm j

l
l
j
j

vl
L
vj

j 1,...,

j l

(4.13)
Now suppose that the decision variables for suppliers and manufacturer are fixed.
Then the retailers problem of finding the optimal replenishment cycle becomes:
98

Chapter 4

min U rl
krl

TDrl
2krl

hrl

krl Orl
T

(4.14)

The value of krl that minimize U rl is by the smallest krl * that satisfies (Palar
and Wang 1994):

krl krl 1
*

T 2 hrl Drl
2Orl

krl * krl * 1 .

(4.15)

The optimal krl can be expressed as (Viswanathan and Wang 2003):

2T 2 hrl Drl
krl * 1 1
Orl

/ 2 .

(4.16)

Here, we define a as the largest integer no larger than a.

We then consider the optimal value of rl . From constraints (4.6) and (4.7), we
can obtain lower bound and the upper bound of rl :

Grl max 0,
Arl Pl erlj Grj Gm j s jv cs jv cm j
j
j

j 1,..., L
vj
erll

j l

Gml slvl cslvl cml
vl

(4.17)

1
Grl
erll

A
erlj Grj Gm j s jv cs jv cm j

rl
j
j

j 1,..., L
vj

Gml slvl cslvl cml .

vl

(4.18)

Substituted (4.13) into (4.2), we can see that Z rl is a quadratic function of Grl .
Because the second derivative of Z rl with respect to Grl is negative, we have:
2 Z rl
Grl 2

2erll 0 .

(4.19)

Thus, Z rl is a concave function of Grl .

99

Chapter 4

Set the first derivative of Z rl with respect to Grl equal to zero. Then Grl can
be obtained as:

Grl

Cl hrl T
,

2erll 4krl

(4.20)

where Cl Arl erll Gml slv cslv cml

l
l
vl

erlj Grj Gm j s jv j cs jv j cm j
L
vj
j 1,...,

j l

If Grl obtained from (4.20) is in the interval of Grl , Grl , it is obviously the

optimal reaction Grl * of the retailer. Otherwise, we have to substitute the bounds
(4.17) and (4.18) into (4.2), the bound that provides higher payoff is the optimal
reaction Grl * .
Proposition 4.1: a) When a retail market becomes less sensitive to his retail price, the
retailer will take opportunity to enhance his retail price to maximize his profit.
b) When a product becomes less substitutable by some other product, this product
retailer will reduce his retail price to maximize his profit.
Proof. If the optimal Grl * is achieved within the interval of Grl , Grl , we have:

Grl
erll
Grl
erlj

Arl

erlj Grj Gm j s jv cs jv cm j
j
j

j 1,..., L
vj

j l

2erll 2

Grj Gm j s jv cs jv cm j

vj

2erll

0.

Otherwise, if Grl * is arrived at the boundary, we have:

1
0, or
A

e
Grj Gm j s jv j cs jv j cm j

r
l
r
lj
erll
erll 2 l
j 1,..., L
vj

j l

Grl

100

Chapter 4

1
2 Arl erlj Grj Gm j s jv cs jv cm j
j
j

erll
erll
j 1,..., L
vj

j l

Grl

Grl
erlj
Grl
erlj

0, or

1
erll

1
erll

Grj Gm j s jv j cs jv j cm j
vj

Grj Gm j s jv j cs jv j cm j
vj

When pml is fixed,

prl
erll

Grl
erll

. Thus, the proof of Proposition 4.1 is completed.

4.4.1.2 The manufacturers reactions

Assume that the decision variables for the retailers are fixed. The manufacturers
problem of finding the optimal setup interval in this case becomes:

min U m
T

T
Dr
2 l l

1 Dr
1 l hmpl
kr
Pl
l

S
Om
m l .
T
l nml T

(4.21)

2Oml
2U m 2Sm

l n T 3 0 , the optimal T
T 2
T3
ml

Since the second derivative of (4.21),

for the minimum of U m can be derived from:

U m 1
Drl
T
2 l

1 Dr
1 l hmpl
kr
Pl
l

S
Oml
m2
0,
2
T
l nml T

(4.22)

or

T*

Sm
l

1
Dr
2 l l

Oml
nml

1 Dr
1 l hmpl
krl Pl

Obviously, the optimal T * obtained from (4.23) satisfies constraint (4.12).

101

(4.23)

Chapter 4

The manufacturers problem of finding the optimal replenishment cycle for raw
material used for product l:

nm 1 hmslv
Om
l
min U sv l
slvl DlT l
nml
2
nml T
l vl

(4.24)

The optimal nml that minimize U sv can be expressed as follows:

8Oml

nml * 1 1 2
T hmslv slv Drl

l
l
vl

/ 2

(4.25)

The payoff Z m is the quadratic function about Gml . From constraints (4.7) and
(4.11), we can obtain lower and the upper bounds of Gml :

Gml max 0,
Arl Pl erlj Grj Gm j s jv cs jv cm j
j
j

erll
j 1,..., L
vj

Grl slvl cslvl cml
vl

(4.26)

1
Gml
erll

A
erlj Grj Gm j s jv cs jvj cm j

rl
j

j 1,..., L
vj

j l

Grl slvl cslvl cml
vl

(4.27)

m s second derivative about Gml is

eril 2T
2 Zm
2
.
2erll hmpi hmsiv siv
i
i
Gml 2
i 1,..., L
vi
Pi

If

(4.28)

2Zm
0 , the optimal Gml can be obtained from the first order condition of
Gml 2

Zm :

102

Chapter 4

Z m
Drl erll Gml
Gml

T
2

erjl
j 1,.., L
j l

1
1
kr
l

j 1,.., L
j l

erlj Gm j

2erjl Drj

Pj

T
er
2 ll

1
1
kr
l

2erll Drl

Pl

hmpl

hms

nml 1 hmslv
l

hmp j lvl slv 2TDrl erll

e
T
slvl rll
l

P
2
vl
l

hms

nm j 1 hms jv
jv j
j
2

s jv j TDrj erjl
s jv j erjl T

P
2
j 1,.., L v j
j
j l

(4.29)
Substitute (4.13) into (4.29), we have:

Bl1 Bl 2Ql Bl 3' X l

Gml
erll erll Bl 2 Bl 3

(4.30)

where
Bl1

erjl T

2
h

mp j
ms jv j s jv j Ar j er jj p j er ji pi ;

j 1,..., L Pj
vj
i j ,l

j l

Bl 2 1

erll T

2
hmpl hmslvl slvl ;
Pl
vl

erjl 2T

2
Bl 3
hmp j hms jv j s jv j ;
j 1,..., L Pj
vj

j l

Bl 3' Bl 3 Grl slv cslv cml ;

l
l
vl

Ql Arl erll Grl slv cslv cml erlj Grj Gm j s jv cs jv cm j

l
l
j
j

vl
L
vj

j 1,...,

j l

Xl

hmp
erjl Gm j i
Pi
j 1,..., L
i

j l

1
1
kr
i

nm 1
hms jv s jv
i
j
j
v
2
i

'
er , i l
eril T , eril ' il
.

eril i l

If Gml obtained from (4.30) is in the interval of Gml , Gml , it is the optimal

reaction Gml * of the manufacturer. Otherwise, Z m reaches its maximal value when
103

Chapter 4

Gml is at its upper bound or lower bound. The bound that provides higher payoff is
the optimal reaction Gml * .
2Zm
0 , we also have to find the bound that provides maximal value of Z m .
Gml 2

If

That is the optimal reaction Gml * .

4.4.2 Algorithm for the equilibrium
Since the MR game is partially sequential, the technique of backward induction
can be used to find a sub-game perfect equilibrium (Basar and Olsder 1982).
Therefore, we will try to solve the Nash equilibrium of the RS game first, and then
solve the Nash equilibrium of the whole MR game. We present the following solution
algorithm:
Step 0. Initialize x(0) xm(0) , xr (0) in the strategy profile set X .
Step 1. Denote x m (0) as the strategy profile of all the chain members in x (0) except
for

the

manufacturer.

Fixed

xm* Gm1 * ,..., GmL * , T * , nm1 * ,..., nmL *

x m (0)

find

the

optimal

reaction

(as Section 4.4.1.2) to optimize the payoff

function Z m in his strategy set X m .

Step 2. Denote xr l (0) as the strategy profile of all the chain members in x (0) except
for retailer l. For each retailer l, fixed xr l (0) , find the optimal reaction

xrl * Grl * , krl *

(as Section 4.4.1.1) to optimize the retailer ls payoff function Z rl

in his strategy set X rl . If

xr * xr (0) 1 , the equilibrium of the RS game xr *

obtained, Go step 3. Otherwise, xr (0) xr * , repeat step 2.

Step 3. x* xm* , xr * . If

x* x(0) 2 , the equilibrium of the MR game x*

104

Chapter 4

achieved. Output the optimal results and stop. Otherwise, x(0) x* , go step 1.
(: 1 , 2 are very small positive numbers)

4.5 NUMERICAL EXAMPLE AND SENSITIVITY ANALYSIS

The primary purchase of this section is to demonstrate the results of the proposed
dynamic non-cooperative game and its solution algorithm by a numerical example. As
an illustration, a case of three retailers and three raw material suppliers is discussed.
The related input parameters for the base example are based on the suggestions from
other researchers (Lu 1995, Woo et al 2001, Prafulla et al. 2006). For example, the
holding cost per unit final product at any retailer should be higher than the
manufacturers. The manufacturers setup cost should be much larger than any
ordering cost. By applying the above solution procedure in section 4.4.2, the optimal
decisions of the manufacturer and all the retailers are presented in Table 4.1.
Sensitivity analysis has been conducted with the model for parameters in three-groups:
market related parameters (e.g., Arl , erll , erij ), production related parameters (e.g.,

Pl , S m , cml ), and holding cost parameters (e.g. , hmpl , hrl ). Both the retailers and
products have the same cost structure. Without loss of generality, we only consider the
related parameters for retailer 1 and product 1. The results of the sensitivity analysis
are also shown in Table 4.1. Several interesting findings are summarized as follows:
Firstly, the manufacturer and the retailer could benefit from the increase of his
market scale / product substitutability or the decrease of price elasticity, while it is

105

Chapter 4

opposite for the other retailers. A larger Ar1 implies greater market scale for product
1. As er11 decreases, the retailer 1s market demand becomes less sensitive to his
retail price. Product 2 and product 1 are more substitutable with a larger er12 .
Referred to Table 4.1, under these cases, the manufacturer would like to produce more
product 1 for a higher profit because of higher price and larger demand for retailer 1
and cut down the production for other products. The increase prices and less market
demand result in the profit decrease for the other retailers.
Secondly, when the manufacturers production cost for one product increases, the
retailer of this product will have a most significant profit decrease, but the other
retailers profits will increase. A larger cml implies a higher cost per unit product l.
Thus, its retailer will have the larger decrease. The manufacturer will seek for higher
profits from other products to cover the loss brought by this product. Thus, he will
reduce his wholesale prices to stimulate their market demand. The other retailers
could benefit from the lower prices charged by the manufacturer. Besides, for the
competition between these products, the increase in price of this product also
enhances the market demand of the other products. So the other retailers profits will
increase.
Thirdly, the manufacturers production capacity, setup cost and holding cost
parameters have significantly influence on the manufacturers setup time interval,
while their influence on the chain members price decisions and profits is minor. The
pricing decisions and profits are less sensitive to Pl , S m , hmpl , hrl , because they

106

Chapter 4

affect mainly the inventory decisions of the manufacturer.

Fourthly, the increase of the manufacturers holding cost for final products might
increases the retailers profits. As hmpl increases, the manufacturer is willing to
shorten his setup time interval and meanwhile, lower the wholesale prices of his
products to offset the impact of higher inventory cost, which will benefit the retailers.
Lastly, the increase of retailers holding cost will lengthen the manufacturers
setup time interval but shorten his own replenishment cycle time. This is because the
retailer would reduce his inventory stock through a higher order frequency. However,
the manufacturer would not like to share much of the retailers inventory. So he would
like to lengthen his setup time interval.

4.6 SUMMARY

We have considered coordination of pricing and inventory decisions in a

seller-buyer supply chain composed of one single manufacturer and multiple
competing retailers in this chapter. The manufacturer and the retailers determine their
optimal marketing (profit margin) and inventory policies (replenishment cycle) to
maximize their individual net profits. This coordination problem is modeled as a
dynamic non-cooperative game. We use both analytical and computational methods
for the derivative of the optimal decisions of all the chain members. A numerical
example has been studied to examine the game model and solution algorithm.
Sensitivity analysis has been conducted on different parameters. The numerical results

107

Chapter 4

show that: (a) when one retailers market scale becomes larger or more sensitive to his
price or his product becomes more substitutable for other product, his profit will be
increased, while the other retailers profits will decrease; (b) the increase of retailers
holding cost will lengthen the manufacturers setup time interval, while shorten the
retailers replenishment cycle time; (c) the change of manufacturers production
capacity or setup cost has mainly impact on the manufacturers setup time interval; (d)
the increase of the manufacturers holding cost may increase the profits of the
retailers.

108

Table 4.1. Results for manufacturer and retailers

(a) The results for manufacturer

pm1

pm2

pm3

nm1

nm2

nm3

Z m (107)

77.69

104.79

175.27

0.0714

1.2554

180,000

72.85

99.22

156.13

0.0741

1.1400

220,000

82.52

110.36

194.40

0.0698

1.3593

1,500
1,700
80
200
5
25
200
1,000
150,000
800,000
1
7
2.5
18

81.34
74.47
76.91
81.69
71.54
83.83
77.69
77.69
77.67
77.69
77.69
77.67
77.70
77.67

105.96
103.64
103.63
110.91
109.23
100.35
104.79
104.80
104.79
104.79
104.79
104.78
104.80
104.78

179.17
171.41
172.19
191.14
190.52
160.02
175.27
175.28
175.26
175.27
175.28
175.25
175.31
175.22

0.0711
0.0717
0.0716
0.0702
0.0702
0.0733
0.0447
0.1029
0.0782
0.0707
0.0857
0.0434
0.0691
0.0733

1
1
1
1
1
1
1
1
1
1
1
1
1
1

1
1
1
1
1
1
1
1
1
1
1
1
1
1

1
1
1
1
1
1
2
1
1
1
1
2
1
1

1.2946
1.2193
1.2365
1.3475
1.3341
1.1689
1.2582
1.2500
1.2542
1.2554
1.2553
1.2552
1.2565
1.2538

Base example

Ar1

er11
er12

cm1
109

Sm
P1
hmp1

hr1

Chapter 4

(b) The results for retailers

Base example

Ar1

180,000
220,000
1,500
1,700
80
200
5
25
200
1,000
150,000
800,000
1
7
2.5
18

er11
er12
cm1
110

Sm
P1
hmp1
hr1

pr1

pr2

pr3

kr1

kr2

kr3

111.69
102.73
120.66
118.45
105.74
110.27
119.02
108.84
114.55
111.70
111.69
111.68
111.70
111.70
111.69
111.69
111.70

158.00
154.46
161.54
158.92
157.11
157.30
161.68
160.44
155.55
158.00
158.00
158.00
158.00
158.00
157.99
158.01
157.99

287.25
277.16
297.33
289.50
285.06
285.62
295.65
294.89
279.61
287.25
287.25
287.24
287.25
287.25
287.24
287.27
287.23

5
5
6
5
5
5
5
5
5
3
8
6
5
6
3
3
8

6
7
6
6
6
6
6
6
6
4
9
7
6
7
4
6
6

3
4
3
3
3
3
3
3
4
2
5
4
3
4
2
3
3

Dr2

Dr1
4

Dr3
4

(10 )
5.4323
4.7712
6.0942
5.5587
5.3061
5.3288
5.9642
5.9591
4.9055
5.4320
5.4328
5.4342
5.4321
5.4319
5.4334
5.4338
5.4311

Z r1
4

(10 )
7.4441
7.7291
7.1595
7.4093
7.4820
7.5087
7.1028
7.1649
7.7233
7.4445
7.4439
7.4443
7.4441
7.4439
7.4447
7.4434
7.4449

(10 )
4.4770
4.8394
4.1150
4.4109
4.5435
4.5350
4.1779
4.1726
4.7816
4.4771
4.4770
4.4774
4.4769
4.4770
4.4774
4.4761
4.4778

Z r2
6

Z r3
6

(10 )
1.7688
1.3538
2.2208
1.9830
1.5819
1.7003
2.1417
2.1380
1.4338
1.8170
1.6675
1.7454
1.7694
1.7343
1.8186
1.8317
1.5603

(10 )
3.8641
4.1482
3.5727
3.8281
3.9036
3.9320
3.5151
3.5778
4.1605
3.9200
3.7527
3.8376
3.8650
3.8256
3.9213
3.8664
3.8625

(10 )
4.9883
5.8203
4.2127
4.8419
5.1379
5.1189
4.3430
4.3320
5.6820
5.0022
4.9544
4.9779
4.9884
4.9737
5.0031
4.9872
4.9899

Initialization for base example: hs 0.01 , hs 0.008 , hs 0.002 , hms hms hms 0.05 , hms 0.02 , hms 0.04 , hms 0.02 , hmp 2 , hmp 1 ,
1

1 v1

2 v1

3 v1

1 v2

1 v3

hmp 0.8 , cs1v 0.93 , cs1v cs2v 2 , cs1v cs2 6 , cs 4 , cm 15 , cm 25 , cm 10 , s 1 , s s 2 , s 3 , s 5 , s

3

P2 250000 ,

P3 100000 ,

hr 8 ,
1

3 v1

v2

hr 6 , hr 5 , Ar 200000 ,
2

1 v1

1 v2

3 v1

Ar 150000 ,

Ar 250000 ,

1 v3

er 1600 ,
11

2 v1

32

33

4 , S m 500 , P1 500000 ,

12

21

13

22

Chapter 4

31

2 v2

er er 100 , er 60 , er 1400 ,

er 120 , er 30 , er 40 , er 400 , Or 40 , Or Or 30 , Om 12 , Om 23 , Om 13 , Rm Rr1 Rr2 Rr3 1000 .

23

2 v2

CHAPTER 5
GAME-THEORETIC COORDINATION OF PRICING AND INVENTORY
DECISIONS IN A MULTI-LEVEL SUPPLY CHAIN

5.1 INTRODUCTION

In recent years, supplier-manufacturer relationship has received great attention in

supply chain management. As indicated by Morgan (1996), supplier involvement in
critical supply chain activities (such as product development, pricing, etc.) enables
manufacturers to achieve the four dimensions of customer satisfaction: competitive
pricing, product quality, product variety and delivery service.
In the previous chapter, we apply a dynamic game structure to study the
coordinating pricing and inventory decisions of supply chain problem (CPISC) in a
two-echelon supply chain manufacturer tier and retailer tier. However, the goal of
supply chain management is to optimize the activities of different individual entities
of the entire supply chain (Lambert et al. 1998; Simchi-Levi et al. 2000). This chapter
extends the CPISC problem investigated in last chapter to a multi-level supply chain
consisting of multiple suppliers, a single manufacturer and multiple retailers.
Studies on CPISC problem have mainly focused on coordination of individual
entities or two-stage channels. In reality, a supply chain usually consists of multiple
firms (suppliers, manufacturers, retailers, etc.). Munson and Rosenblatt (2001)
suggest quantity discount can be used as a mechanism to coordinate inventory in a

111

Chapter 5

three-level

supply

chain

(supplier-manufacturer-retailer).

Giannoccaro

and

Pontrandolfo (2004) propose a contract model based on the revenue sharing

mechanism to coordinate a decentralized control three-stage supply chain. In their
research, the authors focus on the multi-level supply chain with an echelon of one
firm only. The works that consider more than one firms at each echelon are few.
Moutaz (2003) develops a three inventory coordination mechanisms for the supply
chain composed of a single supplier, three manufacturers and two retailers per
manufacturer. Jabber and Goyal (2008) consider coordination of order quantity in a
multiple suppliers, a single vendor and multiple buyers supply chain. These studies
focus on coordination of inventory policies in three-level supply chains with multiple
firms at each stage. However, pricing as an important marketing strategy affecting the
profitability of the enterprises (Elmaghraby and Keskinocak 20003) should also be
involved in supply chain coordination.
We investigate the CPISC problem in a three-level supply chain, where the
manufacturer purchases different types of raw materials from his suppliers. Sole
sourcing strategy is adopted between the manufacturer and the suppliers. Then the
manufacturer uses the raw materials to produce different products for different
independent retailers with limited production capacity. For simplicity, we assume that
the retailers markets are independent of each other, for example, they can be
geographical dispersed. Thus, there is no competition for customers between the
retailers. In this supply chain, all the chain members are rational and determine their

112

Chapter 5

pricing and replenishment decisions to maximize their own profits non-cooperatively.

Our main objective in this chapter is to determine pricing and inventory decisions
for individual non-cooperative firms of the multi-level supply chain to maximize their
profits. To be consistent, we retain the research questions raised for CPISC problem in
Chapter 4 from the manufacturers and retailers perspectives. More specifically, the
questions faced by the retailers are:
(Q5.1) what are the optimal retail prices for products sold to different markets?
(Q5.2) how frequent should the product procurement take place?
For the manufacturer, he has to answer:
(Q5.3) what are the optimal product wholesale prices for different retailers?
(Q5.4) how frequent should the production process take place?
As to each supplier, we are particularly interested in:
(Q5.5) what is the optimal raw material price charged for the manufacturer?
(Q5.6) how frequent should the raw material procurement take place?
In this chapter, the CPISC problem is described as a three-level nested Nash game
with respect to the overall supply chain. The suppliers formulate the bottom-level
Nash game and as a whole play the middle-level Nash game with the manufacturer.
Last, the suppliers and the manufacturer being a group formulate the top-level Nash
game with the retailers. The three-level nested Nash game settles to an equilibrium
solution such that any chain member cannot improve his profits by acting unilaterally
without degrading the performance of other players. We propose both analytical and

113

Chapter 5

computational methods to solve this nested Nash game.

The nested Nash game model and the proposed solution algorithm constitute a
powerful decision support for solving the CPISC problem. Its use is demonstrated and
tested through a numerical example. Several sensitivity analyses are conducted to
examine the influence of the market, production and raw material related parameters
on decisions and profits of all the chain members.
This chapter is organized as follows. The next section gives the problem
description and notations to be used. Section 5.3 develops the three-level nested Nash
game model for this CPISC problem. Section 5.4 proposes the analytical and
computational methods used to solve the CPISC problem in Section 5.3. In Section
5.5, a numerical study and corresponding sensitivity analysis for some selected
parameters have been presented. Lastly, Section 5.6 concludes this chapter.

5.2 PROBLEM STATEMENT AND NOTATIONS

5.2.1 Problem description and assumptions

In the three-level supply chain, we consider the retailers facing the customer
demands of different products, which can be produced by the manufacturer with
different raw materials purchased from the suppliers. These non-cooperative suppliers
reach an equilibrium and as a whole negotiate with the manufacturer on their pricing
and inventory decisions to maximize their own profits. After the suppliers and the
manufacturer reach an agreement, the manufacturer will purchase these raw materials

114

Chapter 5

to produce different products for the retailers. Negotiation will also be conducted
between the manufacturer and the retailers on their pricing and inventory decisions.
When an agreement is reached between them, the retailers will purchase these
products and then distribute to their customers. There is no competition between the
retailers, so they do not need to negotiate for the decisions. We then give the
following assumptions of this chapter:
(A5.1): Each retailer only sells one type of product. The retailers markets are
assumed to be independent of each other. The annual demand function for each
retailer is the decreasing and convex function with respect to his own retail price.
(A5.2): Sole sourcing strategy is adopted between suppliers and manufacturer. That is
to say, each supplier provides one type of raw materials to the manufacturer and the
manufacturer purchases one type of raw material from only one supplier.
(A5.3): The integer multipliers mechanism (Moutaz 2003) for replenishment is
adopted. That is, each suppliers replenishment cycle time is an integer multiplier of
the manufacturers setup time interval which is also the integer multipliers of all the
retailers replenishment cycle.
(A5.4): The inventory of the raw materials for the manufacturer only occurs when
production is set up (Moutaz 2003). Under this assumption, the manufacturer has no
need to keep raw materials when there is no use.
(A5.5): Shortage are not permitted, hence the annual production capacity is greater
than or equal to the total annual market demand (Esmaeili et al. 2009).

115

Chapter 5

(A5.6): The suppliers, the manufacturer and the retailers are rational decision makers
and have equal market power.
5.2.2 Notations

All the input parameters and variables used in our models will be stated as follows.
Assume the following relevant parameters for the retailer:
L: Total number of retailers

rl : Index of retailer l

Arl : A constant in the demand function of retailer l, which represents his market
scale

erl : Coefficient of the products price elasticity of the demand for retailer l
prl : Retail price charged to the customer by retailer l
Drl : Retailer ls annual demand
Orl : Ordering processing cost for retailer l per order of product l
Rrl : Retailer ls annual fixed costs for the facilities and organization to carry this
product

X rl : Strategy set of retailer l. xrl X rl is his strategy

Z rl : Payoff function of retailer l

The retailers decision variables are:

Grl : Retailer ls profit margin

krl : The integer divisor used to determine the replenishment cycle of retailer l
The manufacturers relevant parameters are:
116

Chapter 5

m: Index of manufacturer

hmpl : Holding costs per unit of product l inventory

hmrs : Holding costs per unit of raw material purchased from supplier v
v

S m : Setup cost per production

Om : Ordering processing cost per order of raw materials
Pl : Annual production capacity product l, which is a known constant
Rm : Manufacturers annual fixed costs for the facilities and organization for the
production of this product

pml : Wholesale price charged by the manufacturer to the retailer l

cml : Production cost per unit product l for the manufacturer
X m : Strategy set of the manufacturer. xm X m is his strategy

Z m : Payoff function of the manufacturer

The manufacturers decision variables are:

Gml : Manufacturers profit margin for product l

T : Manufacturers setup time interval

The relevant parameters for the supplier are:

V: Total number of suppliers

sv : Index of supplier v, v=1,2,,V

hsv : Holding costs per unit of raw material inventory for supplier v
csv : Raw material cost paid by supplier v
Rsv : Supplier vs annual fixed costs for the facilities and organization to carry the

117

Chapter 5

raw material

Osv : Order processing cost for supplier v per order

s : Usage of supplier vs raw material to produce a unit product l

v

psv : Raw material price charged by supplier v to the manufacturer

X sv : Strategy set of supplier v. xsv X sv is his strategy
Z sv : Payoff function of supplier v
The suppliers decision variables are:

Gsv : Supplier vs profit margin

K sv : The integer multiplier used to determine the replenishment cycle of supplier

v

5.3 MODEL FORMULATION

5.3.1 Game scheme

We model the CPISC problem as a three-level nested Nash game with V+ L+1
players, i.e., V suppliers, one manufacturer and L retailers, which is an extension of
the two-level nested Nash game discussed in Chapter 4. Each supplier controls
strategy set X sv (v=1,2,,V) to maximize his payoff function Z sv . A strategy

xsv X sv includes profit margin Gsv and replenishment decisions

K sv . The

manufacturer controls the strategy set X m to maximize his payoff function Z m . His
strategy xm X m consists of profit margin Gml for different product l (l=1,2,L)
and setup time interval T. Each retailer controls the strategy set X rl , whose strategy

118

Chapter 5

xrl X rl is composed of profit margin Grl and replenishment decisions krl , to

maximize his payoff function Z rl .
In our game framework, firstly, the V suppliers formulate the bottom-level Nash
game (SS game). Given the manufacturers and the retailers strategies, each
suppliers strategy xsv X sv varies with the change of the other suppliers decisions
to maximize his payoff function Z sv . When none of them would like to alter their
decisions, the bottom-level Nash equilibrium is obtained. Secondly, the suppliers in
equilibrium being a player formulate the middle-level Nash game with the
manufacturer (MS game). In this game, given the retailers decisions, the suppliers

adjust their decisions xs xs1 ,..., xsV

with the change of the manufacturers

decisions, while the manufacturer varies his strategy xm X m as the suppliers

decisions changing until none of them could improve his payoff function by
unilaterally altering his decisions. Thus, the middle-level Nash equilibrium achieves.
The top-level Nash game (RMS game) is also played between the suppliers and the
manufacturer in equilibrium and all the retailers. Each retailers strategy xrl X rl
varies with the change of the suppliers and the manufacturers decisions and the other
retailers decisions. The suppliers and the manufacturer also adjust their equilibrium
strategies

xs , xm

with the change of the retailers decisions. The process continues

until the suppliers, the manufacturer and the retailers cannot increase their payoffs by
changing their decisions. That is the top-level Nash equilibrium reaches.
As Chapter 4 describes, the proposed RMS game is a dynamic single-act game,

119

Chapter 5

where the retailers strategies xrl

( l 1, 2,..., L ) in RMS game affect the

manufacturers strategy xm and the suppliers strategies xsv ( v 1, 2,...,V ) in MS

game, and the decisions from MS game also affect those decisions in the RMS game.
At the same time, the manufacturers strategies in MS game also affect the suppliers
strategies xsv in SS game and vice versa. Through the dynamic interactions between
RMS game, MS game and SS game, individual suppliers, manufacturer and retailers
could determine optimal decisions to maximize their own payoffs. This game is
partially dynamic, because it is a simultaneous game between the suppliers, the
manufacturer and the retailers. Each player acts only once and move simultaneously
(i.e., statically) in the game. That is for the suppliers, the manufacturer and the
retailers, each of them determines their optimal decisions simultaneously. The whole
RMS game described in this chapter is also a non-cooperative game with complete
information where each player knows the other players strategy sets but they are not
involved in any binding agreements. Figure 5.1. shows the three-level nested Nash
game structure.

Supplier

Manufacturer

Retailer

The bottom-level
SS game

The middle-level
MS game

The top-level
RMS game

Figure 5.1. Three-level nested Nash game structure

120

Chapter 5

5.3.2 The retailers model

We first consider the payoff function Z rl for the retailers. The retailers objective
is to maximize his net profit by optimizing his strategy xrl , including profit margin

Grl and replenishment decision krl .

The demand at retailer l (l=1,2,L) is described by a general demand function of
the retail price. We employ the linear demand function in McGuire and Staelin (1983)
and Jeuland and Shugan (1988):

D( prl ) Arl erl prl ,

(5.1)

where Arl is a constant and erl is coefficient of the products demand elasticity. It
can be seen that this demand function is downward sloping and convex functions with
respect to prl .
As indicated in the third point of the assumption in section 5.2.1, the integer
multipliers mechanism is employed between the manufacturer and the retailers. Since
the setup time interval for the manufacturer is assumed to be T, the replenishment
cycle for retailer l is

cost is

TDrl
2krl

T
. krl should be a positive integer. Thus, the annual holding
k rl

hrl (see Figure 5.2(a)) and the ordering process cost is

121

Orl krl
T

Chapter 5

Inventory level

TDrl
krl
Order point

T
krl

(a) Retailer ls inventory level of final product

Time

Inventory level

Level of raw material u

s TDr
lu

Level of raw material v

s TDr
lv

Level of product l

TDrl
Pl

T
krl

Time

(b) Manufacturers inventory level for product l

Inventory level

Ksv T slv Drl

l

Time

Ksv T
(c) Supplier vs inventory level of raw material

Figure 5.2. Inventory levels of retailer, manufacturer, supplier

122

Chapter 5

The retailer l faces the holding cost, the ordering cost and an annual fixed cost and
the retailer ls objective function is expressed by the following equation:

max Z r Gr Dr
krl ,Grl

TDrl
2krl

hrl

Orl krl
T

Rrl ,

(5.2)

Subject to

krl {1 , 2 , 3 ,, . . . }

(5.3)

Grl prl pml ,

(5.4)

Drl Arl erl p,rl

(5.5)

Grl 0 ,

(5.6)

0 Drl Pl .

(5.7)

Constraint (5.3) gives the value of the divisor used to determine the retailer ls
replenishment cycle time. Constraint (5.4) indicates the relationship between the
prices (the retail price and the wholesale price) and retailer ls profit margin.
Constraint (5.6) ensures that the value of Grl is nonnegative. Constraint (5.7) gives
the bounds of the annual demand, which cannot exceed the annual production
capacity of the product.
5.3.3 The manufacturers model

The manufacturers objective is to determine his strategy xm , composed of the

profit margins for all the products Gml and the setup time interval for production T,
to maximize his net profit.
The manufacturer faces annual holding costs, setup and ordering costs, and an

123

Chapter 5

annual fixed cost. The annual holding cost for the manufacturer is composed of two
parts: the cost of holding raw materials used to convert to products, the cost of
holding products. During the production portion, the average inventory of raw
material v used for product l is

s TDr
lv

. The production time in a year is

Drl
Pl

During the non-production portion of the cycle, the raw materials inventory drops to
zero and the holding cost is zero according to our assumption. Hence, the annual
holding cost for raw material v is

ls is given by

T
Dr
2 l

s T Dr
lv

2 Pl

hmrs . The annual inventory for product

v

1 Dr
1 l (Lu 1995). The behavior of the inventory level
kr
Pl
l

for the manufacturer is illustrated as Figure 5.2(b). The setup cost S m and ordering
cost Om occur at the beginning of each production. Thus, we can derive the
manufacturers payoff function Z m :

max

Gm1 ,...,Gml ,...,GmL ,T

T
Z m Gml Drl Drl
2 l
l

1 Dr
1 l hmpl
kr
Pl
l

s T Drl
lv
l v
2 Pl

hmrs
v

Sm Om
Rm
T

(5.8)
Subject to

Gml pml slv psv cml , for each l=1,2,L

(5.9)

Gml 0 , for each l=1,2,L

(5.10)

T 0.

(5.11)

Constraint (5.9) gives the relationship between the price (the wholesale price and
the raw material price) and the manufacturers profit margin. Constraint (5.10) and

124

Chapter 5

(5.11) ensure that the values of Gml and T are nonnegative.

5.3.4 The suppliers model

Each suppliers problem is to determine an optimal strategy xsv (including

replenishment decisions K sv and profit margin Gsv ) to maximize his net profit.
According to the third point of the assumption in section 5.2.1, the integer
multipliers mechanism is adopted between the suppliers and the manufacturer. So the
replenishment cycle time for supplier v is K sv T . The raw material inventory drops
every T year by T slv Drl starting from ( K sv 1)T slv Drl as Figure 5.2(c)
l

shows. Therefore, the holding cost is:

( K sv 1)T slv Drl (Thsv ) + ( K sv 2)T slv Drl (Thsv ) ++ T slv Drl (Thsv ) ,
l

( K sv 1)T slv Drl

l

which is equal to

hsv . The supplier faces holding costs, ordering

costs, and an annual fixed cost. Thus, the supplier vs payoff function Z sv is:

max Z s Gs

K sv ,Gsv

s Dr
lv

( K sv 1)T slv Drl

hsv

Osv
K sv T

Rsv

(5.12)

Subject to

K sv 1 , 2 , 3, ,. . .

(5.13)

Gsv psv csv ,

(5.14)

Gsv 0 .

(5.15)

Constraint (5.13) gives the value of suppliers multiplier used to determine his
replenishment cycle time. Constraint (5.14) indicates the relationship between the raw
material price and the suppliers profit margin. Constraint (5.15) ensures the
125

Chapter 5

non-negativeness of Gsv .
5.3.5 Three-level nested Nash game model

As the game structure shown in Figure 5.1 and the above payoff functions of the
players, we can formulate the three-level nested Nash game model as follows:

M ax Z r , for each l=1,2,,L

krl ,Grl

Subject to
Constraints (5.3)-(5.7)

max Z m
T ,Gml

Subject to
Constraints (5.9)-(5.11)

max Z s , for each v=1,2,,V

K sv ,Gsv

Subject to
Constraints (5.13)-(5.15)
The retailers and suppliers have the same cost structures, so we can use one
uniform formula to represent the payoff function Z rl of each retailer as defined in
(5.2)-(5.7) and one uniform formula for each suppliers payoff function Z sv as in
(5.12)-(5.15). In the above game model, payoff function Z rl and constraints
(5.3)-(5.7) for each retailer l=1,2,,L in this game model formulate the RMS game.
Payoff function Z m and constraints (5.9)-(5.11) formulate the MS game, which is the
model of the manufacturer. The SS game is formulated by payoff function Z sv and
constraints (5.13)-(5.15) for each supplier v=1,2,,V.

5.4 SOLUTION ALGORITHM

Similar to Chapter 4, in this chapter, we also employ analytical theory introduced

126

Chapter 5

in Liu (1998) to compute Nash equilibrium. To determine the three-level nested Nash
equilibrium, we first use analytic method to calculate the best reaction functions of
each player and then build the Nash equilibrium through algorithm procedure.
5.4.1 Reaction functions

5.4.1.1 The retailers reactions

Substituting (5.4), (5.9), (5.14), we can rewrite the retailer ls demand function
(5.5) as:

Drl Arl erl Grl Gml slv Gsv csv cml

v

(5.16)

Suppose that the decision variables for suppliers and manufacturer are fixed. The
retailers problem of finding the optimal replenishment cycle becomes:

min U rl
krl

TDrl
2krl

hrl

krl Orl

(5.17)

The best reaction krl can be expressed as:

2T 2 hrl Drl

krl 1 1
Orl

/ 2 .

(5.18)

a is defined as the largest integer no larger than a.

We then consider the optimal value of Grl . From constraints (5.6) and (5.7), we
can obtain lower bound and the upper bound of Grl :

Ar Pl
Grl max 0, l
Gml slv Gsv csv cml

e
v

r
l

Grl

Gml slv Gsv csv cml

erl
v

Arl

127

(5.19)

(5.20)

Chapter 5

Substituted (5.16) into (5.2), we can see that Z rl is a quadratic function of Grl .
Because the second derivative of Z rl with respect to Grl is negative, we have:
2 Z rl
Grl 2

2erl 0 .

(5.21)

Thus, Z rl is a concave function of Grl .

Set the first derivative of Z rl with respect to Grl equal to zero. Then Grl can
be obtained as:

Grl

Cl hrl T
,

2erl 4krl

(5.22)

where Cl Arl erl Gml slv Gsv csv cml

v

If Grl obtained from (5.22) is in the interval of Grl , Grl , it is obviously the

optimal reaction Grl * of the retailer. Otherwise, we have to substitute the bounds
(5.19) and (5.20) into (5.2), the bound that provides higher profit is the best reaction

Grl * .
5.4.1.2 The manufacturers reactions
Assume that the decision variables for the suppliers and the retailers are fixed. The
manufacturers problem of finding the optimal setup interval in this case becomes:
T
min U m Drl
T
2 l

1 Dr
1 l hmpl
kr
Pl
l

s T Drl
lv
l v
2 Pl

Since the second derivative of (5.23),

hmrs
v

Sm Om
. (5.23)
T

2 Sm Om
2U m

0 , the optimal T
2
T
T3

for the minimum of U m can be derived from:

128

Chapter 5

U m 1
Drl
T
2 l

1 Dr
1 l hmpl
kr
Pl
l

s Drl
lv
l v
2 Pl

hmrs
v

Sm Om
0
T2

(5.24)

(5.25)

or

T*

Sm Om
1
Dr
2 l l

s Drl
lv
l v
2 Pl

1 Dr
1 l hmpl
krl Pl

hmrs

Obviously, the optimal T * obtained from (5.25) satisfies constraint (5.11).

The net profit Z m is the quadratic function about Gml . From constraints (5.7)
and (5.10), we can obtain lower and the upper bounds of Gml :

Ar Pl
Gml max 0, l
Grl slv Gsv csv cml

e
v

r
l

Gml

Grl slv Gsv csv cml

erl
v

Arl

(5.26)

(5.27)

Z m s second derivative about Gml is

2 Zm
1
2erl 1
2
kr
Gml
l

erl 2T
erl T
hmpl slv hmrsv

P
v

2Zm
If
0
Gml 2

(5.28)

(5.29)

the optimal Gml can be obtained from the first order condition of Z m :
Z m
T
Drl erl Gml Drl erl
Gml
2

2erl Drl
s er TDr
hmpl lv l l hmrs 0 (5.30)
v
Pl
Pl
v

1 k1

rl

Substitute (5.16) into (5.30), we have:

hmpl T
1
1
2 krl
Gml
Wml
hmpl T
1 2erl erl T
2
1

s hmr
2 krl
Pl Pl v lv sv
Wml

129

(5.31)

Chapter 5

where Wml

Grl slv Gsv csv cm1 .

erl
v

Arl

If Gml obtained from (5.31) is in the interval of Gml , Gml , it is the optimal

reaction Gml * of the manufacturer. Otherwise, Z m reaches its maximal value when

Gml is at its upper bound or lower bound. The bound that provides higher profit is the
optimal reaction Gml * .
If

2Zm
0 , we also have to find the bound that provides maximal value of Z m .
Gml 2

That is the best reaction Gml * .

5.4.1.3 The suppliers reactions
Lastly, we consider the reaction functions for the suppliers. Suppose that the
decision variables for retailers and manufacturer are fixed. The suppliers problem of
finding the optimal replenishment cycle becomes:
min U sv

( K sv 1)T slv Drl

l

K sv

hsv

Osv
K sv T

(5.32)

The optimal K sv that minimize U sv can be expressed as follows:

K sv

8Osv

1 1 2
T hsv slv Drl

/ 2 .

(5.33)

We then consider supplier vs optimal reaction for Gsv . The second order
condition for Z sv ,
2 Z sv
Gsv

2 slv 2erl 0 .

(5.34)

Thus the necessary condition to maximize the suppliers net profit Z sv is:

130

Chapter 5

Z sv
Gsv

slv Drl Gsv slv 2erl

l

( K sv 1)T slv 2erl

l

hsv 0 .

(5.35)

Substitute (5.16) into (5.35), we can obtain:

slv erl
l

Gsv

u 1,...,V
u v

s Gs s El
lu

lv

( K sv 1)T

2 slv erl
2

hsv ,

(5.36)

where El Arl erl Grl Gml slv csv cml

v

From constraints (5.7) and (5.15), we can obtain the upper bound and lower bound
of Gsv :

Ar Pl
1
Gsv max 0, l

Grl Gml slu Gsu csu slv csv cml

u v

r
s
s
l
lv
lv

Gsv

Arl
erl slv

1
G Gml slu Gsu csu slv csv cml
slv rl
u v

(5.37)

(5.38)

If Gsv obtained from (5.36) is in the interval of Gsv , Gsv , it is the best reaction

Gsv * of supplier v. Otherwise, we have to substitute the bounds (5.37) and (5.38) into
(5.12), the bound that provides higher profit is the optimal reaction Gsv * .
5.4.2 Algorithm for the three-level nested Nash equilibrium

In this algorithm, we try to solve the Nash equilibrium of the SS game first, then
solve the Nash equilibrium of the MS game, and lastly solve the Nash equilibrium of
the RMS game.
Similar

as

sub-section

X m Gm1 , Gm2 ,..., GmL , T

5.3.1,

and X sv Gsv , K sv

we

denote

X rl Grl , krl

as the strategy sets of retailer l,

manufacturer and supplier v, respectively. X s X s1 ... X sV , X r X r1 ... X rL ,

131

Chapter 5

X ms X s X m and X X ms X r are the strategy profile sets of the suppliers, the

retailers, the suppliers and the manufacturer, and all the chain members. We present
the following algorithm for solving the three-level nested Nash game model:
Step 0. Initialize x(0)

(0)

, xm(0) , xr (0)

in the strategy profile set I .

Step 1. Denote xr l (0) as the strategy profile of all the chain members in x (0) except
for retailer l. For each retailer l, fixed xr l (0) , find the optimal reaction

xrl * Grl * , krl *

(as Section 5.4.1.1) to optimize the retailer ls payoff function Z rl

in its strategy set I rl .

Step 2. Denote x m (0) as the strategy profile of all the chain members in x (0) except
for

the

manufacturer.

xm* Gm1 * , Gm2 * ,..., GmL * , T *

Fixed

x m (0)

find

the

optimal

reaction

(as Section 5.4.1.2) to optimize the profit function Z m

in its strategy set I m .

Step 3. Denote xs v (0) as the strategy profile of all the chain members in x (0) except
for supplier v. For each supplier v, fixed xs v (0) , find the optimal reaction

xsv * Gsv * , K sv *

strategy set I sv . If

(as Section 5.4.1.3) to optimize the profit function Z sv in its

xs* xs (0) 1 , the Nash equilibrium of SS game xs* obtained,

Go step 4. Otherwise, xs (0) xs* , repeat step 3.

Step 4. xms* xs* , xm* . If

xms* xms (0) 2 , the Nash equilibrium of MS game,

xms* , obtained, go step 5. Otherwise, xms (0) xms* , go step 2.

Step 5. x* xms* , xr * . If

x* x(0) 3 , the Nash equilibrium of RMS game x*

obtained. Output the optimal results and stop. Otherwise, x(0) x* , go step 1.

132

Chapter 5

(: 1 , 2 , 3 are very small positive numbers)

5.5 NUMERICAL EXAMPLE AND SENSITIVITY ANALYSIS

In this section, we present a simple numerical example to demonstrate the

applicability of the proposed solution procedure to our game model. We consider a
supply chain consisting of three suppliers, one single manufacturer and two retailers.
The manufacturer procures three kinds of raw materials from the three suppliers. Then
the manufacturer uses them to produce two different products and distributes them to
two retailers. These parameters for the example are given as: hs1 0.01 , hs2 0.008 ,
hs3 0.002 , Os1 12 , Os2 23 , Os2 13 , cs1 0.93 , cs2 2 , cs3 6 , hmrs 0.05 ,
1

hmrs 0.02 , hmrs 0.04 , hmp1 0.5 , hmp2 1 , cm1 15 , cm2 25 , s11 1 , s12 2 ,
2

s 3 , s 5 , s 4 , s 2 , Sm 1000 , Om 50 , P1 500000 ,
13

21

22

23

P2 300000 , hr1 1 , hr2 2 , Ar1 200000 , Ar2 250000 , er1 1600 , er2 1400 ,

Or1 40 , Or2 30 . And the fixed cost for all the players are 1000. By applying the
above solution procedure in section 5.4.2, the optimal results for the suppliers, the
manufacturer and the retailers are shown in Table 5.1(a). In order to ensure that our
conclusions are not based purely on the chosen numerical values of the base example,
we also conduct some sensitivity analysis on some parameters, including the market
related parameter, the production related parameter and the raw material related
parameter.

133

Chapter 5

5.5.1 Simulation results

5.5.1.1 Sensitivity analysis for market related parameter

Since the retailers have the same cost structure, without any loss of generality, we
consider the situation that er1 is 120% of the base example. Table 5.1(b) presents the
results. Compared Table 5.1(b) with Table 5.1(a), we can draw:
Firstly, the retailer 1 reduces his price by 13.81%, meanwhile his demand still
reduces by 15.39%. Hence, his profit reduces by 40.76%. The retailer 2s prices
reduce by 0.549%, while his profit rises up by 3.453%.
Moreover, as the increase of er1 , the manufacturer reduces the prices of all his
products and decides to produce larger quantity of product 2. However, the increase
profit brought by product 2 cannot compensate the loss for product 1. Hence, the
manufacturers profit still has a decrease of 4.81%. Besides, the manufacturers setup
interval becomes longer, from 0.2691 to 0.2695.
Lastly, all the suppliers optimal prices and profits are reduced except for supplier
1s, whose price and profit increase from 1.77 and 2.5335105 to 1.99 and 3.2029
105, a raise of 12.43% and 20.9% respectively.
5.5.1.2 Sensitivity analysis for production setup cost
To show the influence of the manufacturers production parameter, we increase
the manufacturers setup cost S m by 50%. The results are given in Table 5.1(c).
Compared with Table 5.1(a), we can see:
First of all, the retailer 1s price increases from 102.20 to 102.28, while retailer 2s

134

Chapter 5

price decreases from 140.28 to 140.10. Despite of that, both of their profits reduce
from 8.0236105 and 18.943105 to 7.8407105 and 18.269105, by 2.28% and
3.56% respectively. The manufacturers profit declines from 7.7953105 to 6.7750
105, which has a most decline of 13.1%.
Secondly, the supplier 1 reduces his raw material price and his profit also has a
slightly decrease. However, for the supplier 2 and supplier 3, they rise up their prices,
but surprisingly both of them have a profit increase of around 1.4%.
In addition, the manufacturers setup interval has a significant increase from
0.2691 to 0.3252, by 17.3%.
5.5.1.3 Sensitivity analysis for raw material cost
Suppose that Supplier 1s raw material cost cs1 has a 50% increase. The results
are shown in Table 5.1(d). Through comparison of the results with those of the base
example, we obtain the following conclusions:
With the increase of cs1 , all the chain members push up their prices accordingly.
The demands for retailer 1 and retailer 2 decrease from 3.6486104, 5.3608104 to
3.6426104 and 5.2531104 . The profits for retailer 1 and retailer 2 decrease from
8.0236105 and 18.943105 to 7.9950105 and 18.149105, by 0.356% and
4.192%. The profit for the manufacturer reduces from 7.7953105 to 7.5513105, a
decrease of 3.13%.
Although cs1 raises up, supplier 1s profit margin remains the same as the base
example and his profit only decreases from 2.5335105 to 2.4945105, a slightly

135

Chapter 5

slopping of 0.263%. However, the supplier 2s and supplier 3s profit margins reduce
and their profits decline by 2.863% and 2.565% respectively.
The setup interval for the manufacturer goes up from 0.2691 to 0.2707. That is the
manufacturer sets up it production less frequently than that of the base example.
Correspondingly, the replenishment cycles for the suppliers and the retailers also get
lengthened.
5.5.2 Managerial implications

Through the three-level nested Nash game model and the numerical example,
some meaningful managerial implications can be drawn:
Firstly, the increase of one retailers price elasticity will reduce his profit, but
benefit other retailer. When er1 increases, the change of the retailer 1s demand is
more sensitive to the change of his retail price compared with the base example. The
retailer 1s profit can be less reduced by lowering his retail price. But his market
demand cannot be increased, which makes the manufacturer seek for higher profit
from other product to fill up the loss deduced by this product / retail market. It is good
news to the other product / retailer, because the manufacturer will lower down his
wholesale price to stimulate this market demand.
Secondly, when the manufacturers setup cost S m increases, the manufacturers
profit decreases more significantly than the retailers, while some suppliers profits
increase. The increase of S m makes the manufacturer produce more product with
higher profit margin (product 2) and reduce the production of lower profitable product
136

Chapter 5

(product 1). The usage of raw materials increases as the change of the manufacturers
production strategy. At the same time, some suppliers bump up their prices, thus
bringing higher profits to them. Different from the manufacturer, the adjusting of the
retailers replenishment strategy krl could partially offset the effects of the increase
of the manufacturers setup time interval. So their profits decrease less than the
manufacturers.
Thirdly, the impact of the increase of one suppliers raw material cost on his own
profit may not as significant as that on the other suppliers. The increase of cs1
makes the supplier 1 raise his raw material price and result in an increase cost in final
products, as well as the decrease in market demands. Hence, the other suppliers will
reduce their prices to keep the market and optimize their individual profits. Supplier 1
has the much lower profit margin than other suppliers, so he will not reduce his profit
margin. Hence, the supplier 1s profit decreases least.
Lastly, when the retailers price elasticity, the manufacturers setup cost S m , or
the suppliers raw material cost increase, the manufacturers setup time interval will
be lengthened. A higher erl or csv results in the total market demands decrease, as
well as a lower inventory consumption rate. The increase of S m makes the
manufacturers cost per production hike up. Hence, the manufacturer has to conduct
his production less frequently.

137

Chapter 5

5.6 SUMMARY

In this chapter, we consider coordination of pricing and replenishment cycle in a

multi-level supply chain composed of multiple suppliers, one single manufacturer and
multiple retailers. The suppliers, the manufacturer and the retailers determine their
optimal pricing and inventory policies to maximize their individual net profits. This
CPISC problem is modeled as a three-level nested Nash game model. In this game,
the suppliers formulate the bottom-level Nash game. The supplier sector formulates
the middle-level Nash game with the manufacturer. Lastly, the suppliers and the
manufacturer as a group play the top-level Nash game with the retailers. We use both
analytical and computational methods for the derivative of the optimal decisions of all
the chain members. A numerical example has been studied to examine the game
model and solution algorithm. Sensitivity analysis has been conducted on market
parameter, production parameter and raw material parameter. The results of the
numerical example also show that: (a) when one retailers market becomes more
sensitive to their price, his profit will be decreased, while the other retailers profit
will increase; (b) the increase of the manufacturers production setup cost will bring
losses to himself and the retailers, but may increase the profits of some suppliers; (c)
the increase of raw material cost causes losses to all the supply chain members.
Surprisingly, the profit of this raw materials supplier may not decrease as significant
as the other suppliers; (d) the setup time interval for the manufacture will be

138

Chapter 5

lengthened as the increase of the retailers price sensitivity, the manufacturers setup
cost or the suppliers raw material cost.
Table 5.1. Results for suppliers, manufacturer and retailers
(a) The base example
Supplier
1
Profit margin
0.84
Cycle time
0.2691
Price
1.77
4
Demand (10 ) 30.453
5
Profit (10 ) 2.5335

Manufacturer

Retailer

Product 1

Product 2

4.15
0.2691
6.15
28.740

9.15
0.2691*2
15.15
21.667

4.87
0.2691
79.37
3.6486

13.26
0.2691
101.96
5.3608

22.83
0.2691/6
102..20
3.6486

38.32
0.2691/11
140.28
5.3608

11.909

19.804

8.0236

18.943

7.7953

(b) er1=1920
Supplier
1
Profit margin
1.06
Cycle time
0.2695
Price
1.99
4
Demand (10 ) 29.908
5
3.2029
Profit (10 )

Manufacturer

Retailer

Product 1

Product 2

3.99
0.2695
5.99
28.298

7.58
0.2695*2
13.58
21.434

2.26
0.2695
71.98
3.0870

14.36
0.2695
100.42
5.4689

16.11
0.2695/6
88.09
3.0870

39.09
0.2695/11
139.51
5.4689

11.181

15.305

4.7535

19.597

7.4206

(c) Sm=1500
Supplier
1
Profit margin
0.83
Cycle time
0.3252
Price
1.76
4
Demand (10 ) 30.563
5
2.5306
Profit (10 )

Manufacturer

Retailer

Product 1

Product 2

4.20
0.3252
6.20
28.812

9.27
0.3252
15.27
21.675

4.58
0.3252
79.55
3.6345

12.47
0.3252
101.61
5.3857

22.74
0.3252/7
102.28
3.6345

38.49
0.3252/14
140.10
5.3857

12.081

20.084

7.8407

18.269

6.7750

(d) cs1=0.93+0.93/2
Supplier
Profit margin
Cycle time
Price
4
Demand (10 )
5
Profit (10 )

Manufacturer

Product 1

0.84
0.2707
2.23
29.908
2.4945

4.09
0.2707
6.09
28.298
11.568

9.01
0.2707*2
15.01
21.434
19.296

139

Product 2

5.00
12.96
0.2707
0.2707
79.44
103.50
3.6426
5.2531
7.5513

Retailer
1

22.79
0.2707/6
102.23
3.6426
7.9950

37.55
0.2707/11
141.05
5.2531
18.149

CHAPTER 6
GAME-THEORETIC COORDINATION OF PRICING AND INVENTORY
DECISIONS CONSIDERING SUPPLIER SELECTION IN A MULTI-LEVEL
SUPPLY CHAIN

6.1 INTRODUCTION

The previous chapter has analyzed CPISC problem in a multi-level supply chain,
composing of multiple suppliers, one manufacturer and multiple retailers. It is widely
acknowledged that supplier selection and evaluation have become a major topic in
production and operations management (Mihaly 1999; Michael and Chong 2001).
Monczka et al. (1997) propose several benefits of supplier selection and integration into
new product development and find that many leading manufacturing firms make
supplier selection decisions early in the product design and planning stage. Studies by
some researchers (Ragatz et al. 1997; Tan 2001; McIvor and Humphreys 2004) further
indicate that supplier involvement into the early stage of product development may lead
manufacturers to choose the best product variants and production processes.
In last few decades, the supplier selection problem has gained great attention in the
context of supply chain management. However, only a few models incorporate the
supplier selection to the procurement lot sizing under deterministic demand conditions,
i.e., the demand quantity is known and fixed (Liao and Rittscher 2007). Degraeve and
Roodhooft (1999) propose a mathematical programming model to simultaneously select

140

Chapter 6

suppliers and to determine the order quantities over a multi-period time horizon.
Ghodsypour and OBrien (2001) build single-objective and multi-objective
mathematical programming models minimizing the total cost of logistics in the process
of the supplier selection, including the aggregate price, the ordering cost and the
inventory costs, subject to capacity, budget, quality and delivery requirements. These
studies incorporate supplier selection decisions with inventory polices to improve the
channel profits. However, marketing strategies are not always involved in their models.
The emphasis of this chapter is placed upon the coordination of supplier selection,
pricing and inventory decisions (CSSPI) in a multi-level supply chain consisting of
multiple alternative suppliers, one manufacturer and multiple retailers. The
manufacturer purchases optional components of certain functionality from his
alternative qualified suppliers to produce a set of products to meet the requirements
from the retailers in different markets. The following research questions are raised for
the retailers, the manufacturer and the suppliers in CSSPI problem:
(Q6.1) what are the optimal retail prices for products sold to different markets?
(Q6.2) how frequent should the product procurement take place for the retailer?
(Q6.3) what are the optimal suppliers and components that should be selected for the
manufacturer?
(Q6.4) what are the optimal product wholesale prices charged by the manufacturer?
(Q6.5) how frequent should the production process take place?
(Q6.6) what are the optimal component prices set by the suppliers?

141

Chapter 6

We describe the CSSPI problem as the same three-level nested Nash game with
respect to the overall supply chain as the precious chapter. The suppliers formulate the
bottom-level Nash game and at the same time as a whole play the middle-level Nash
game with the manufacturer. The suppliers and the manufacturer also being a group
formulate the top-level Nash game with the retailers. Once the whole game settles to an
equilibrium solution, none of the chain members is able to improve its payoff (i.e.,
profits) by acting unilaterally without degrading the performance of other players. We
propose both analytical and computational methods to obtain the Nash equilibrium of
this game.
The game model and the proposed solution algorithm constitute a powerful decision
support for solving the CSSPI problem, which is tested through a numerical example.
The impacts of the market scale parameter, production setup cost and component cost
on the decisions and profits of all the chain members are also investigated. We also
study influence of the component selection strategy which implies that the
manufacturer could choose higher functionality components to replace lower ones for
products.
This chapter is structured as follows. The next section presents the problem
description and some notations. We formulate the mathematical model of the CSSPI
problem in Section 6.3. Section 6.4 proposes the analytical and computational methods
used to solve the CSSPI problem. In Section 6.5, a numerical study and sensitivity
analysis have been presented. We conclude this chapter in Section 6.6.

142

Chapter 6

6.2 PROBLEM DESCRIPTION AND NOTATIONS

In this section, we describe the CSSPI problem using an illustrative supply chain
adapted from Grave and Willems (2005) and Huang et al. (2007) in Figure 6.1. This
application case is concerned with a three-echelon supply chain involving multiple
suppliers, one single manufacturer and multiple retailers. The manufacturer, indicated
by m , designs and customizes a set of products for retailers ( rl , l 1, 2,..., L ) in
different independent market segments. Each retailer is served by one product
customized from the product set. In Figure 6.1, the manufacturer focuses on two
computer products, involving Notebook A and Notebook B, and sells them to the
retailers in two market regions respectively, namely European Union (EU) and North
America (NA).
The architecture for the set of products is finalized which consists of a series of
different functionality elements. The components offering different levels of the same
functionality are grouped together in the substitutable component set (SCS), indexed by
i 1, 2,..., I . Suppose that the components within the same SCS can be ranked in order

of decreasing functionality and higher functionality components can substitute ones

with lower functionality completely, but not vice versa. The lowest functionality
components in all SCSs satisfying the products requirement have also been predefined.
In this research, components selection strategy is referred to as whether to choose
higher functionality components to replace lower ones fixed a priori and what higher
functionality components to choose for the manufacturer.

143

Chapter 6

Suppliers

Manufacturer

Processor

Retailers

Intel Core
2 Duo

s1

Intel Pentium
Dual-Core

s2

LCD Display
SXGA+

s3

Notebook A

r1 in EU Market

Notebook B

r2 in NA Market

WXGA

XGA

s4
Memory

Hard Drive

s5
Miscellaneous
Components

Metal
Housing

s6
Alternative
Supplier

Component

SRS with multiple

components

Platform
Product

Manufacturer

Retailer

Figure 6.1. Coordination of suppliers selection, pricing and inventory

decisions (CSSPI): an illustration

Let N i be the number of components in SCS i . Lij is used to denoted the

component which is the

j th element in the i th SCS, where i 1, 2,..., I and

j 1, 2,..., Ni . In Figure 6.1, the two products share the same architecture, with SCSs,
processor, LCD display, memory, hard drive, miscellaneous components and metal
housing, in sequence. Among them, processor and LCD display have several
component options, ranked in order of decreasing functionality. For instance, in
processor SCS, Intel Pentium Dual-core has been fixed for Notebook B, but the
manufacturer could select higher functionality Intel Core 2Duo to replace it. For those
SCSs, which involve one component only, we do not distinguish SCS or component for
144

Chapter 6

them. For example, for the memory SCS and memory component, we use the same sign
to denote, as Figure 6.1 shows.
All the components are purchased from a fixed number of alternative suppliers
( sv , v 1, 2,...,V ), which are selected through certain criteria, such as quality, service,
etc. For the components as Figure 6.1, there are 6 alternative suppliers. Figure 6.1 also
shows the relationship between the suppliers and the components they provided. For
example, supplier 1 provides Intel Core 2Duo processor and SXGA+ display.
The suppliers, the manufacturer and the retailers are assumed to be rational decision
makers. An immediate question faced by each supplier is how to determine the prices
for the components he sells to maximize his net profit. The manufacturer will have to
determine the setup time interval for production, the wholesale prices, and select the
best suppliers and components to maximize his net profit. The retailers problem will
focus on their replenishment cycles and retail prices for the products. Thus, under this
supply chain circumstance, these competing non-cooperative suppliers reach an
equilibrium on their pricing decisions and as a whole negotiate with the manufacturer
on their pricing, inventory and supplier selection decisions to maximize their own
profits. Negotiation will also be conducted between the manufacturer and the retailers
on their pricing and inventory decisions. After the suppliers, the manufacturer and the
retailers reach an agreement, the manufacturer will purchase these components from the
suppliers to produce different products for the retailers.
We then give some other assumptions used for building the mathematical model in

145

Chapter 6

the next section.

(A6.1): The integer multipliers mechanism (Moutaz 2003) for replenishment is adopted
between the manufacturer and the retailers. That is, the manufacturers setup time
interval is integer multipliers of the replenishment cycle time of the retailers.
(A6.2): The suppliers and the manufacturer use vendor managed inventory (VMI)
system (Simchi-Livi et al. 2000; Tyan and Wee 2003) to replenish their components.
This inventory system has been adopted by some industries for years (e.g., Wal-Mart,
Procter&Gamble (P&G), Dell, etc.). Under this assumption, the components
inventories for the manufacturer and the suppliers are at the side of their upstream
suppliers and their corresponding inventory holding costs are also borne by their
upstream suppliers (Birendra and Srinivasan 2004).
(A6.3): Single sourcing strategy (Tullous and Utrecht 1992) is adopted between
supplies and manufacturer. Thus, the manufacturer purchases one type of component
from only one supplier.
(A6.4): Either all or none the demand of a component is replaced by the lower order
component in the same SCS.
(A6.5): Shortage are not permitted, hence the annual production capacity is greater than
or equal to the total annual market demand (Esmaeili et al. 2009).

146

Chapter 6

6.3 THE MATHEMATICAL MODEL

6.3.1 Game scheme

The CSSPI problem is also modeled as a three-level nested Nash game with
V L 1 players, i.e., V suppliers, one manufacturer and L retailers. Each supplier

controls his strategy set X sv ( v 1, 2,...,V ) to maximize his payoff function Z sv . A

strategy xsv X sv includes pricing decisions for the components supplied by the
supplier. The manufacturer controls the strategy set X m to maximize his payoff
function Z m . His strategy xm X m consists of setup time interval, profit margin for
each product, and suppliers and components selection decisions. Each retailer controls
his strategy set X rl , whose strategy xrl X rl is composed of replenishment decisions
and profit margin, to maximize his payoff function Z rl .
In our game framework, the suppliers formulate the bottom-level Nash game (SS
game) and at the same time as a group formulate the middle-level Nash game (MS
game) with the manufacturer. Also, the top-level Nash game (RMS game) is played
between the suppliers and the manufacturer as a whole sector and all the retailers. As
introduced in Chapter 5, the whole RMS game in this chapter is also a dynamic
single-act game and a non-cooperative game with complete information where each
player knows the other players strategy sets but they are not involved in any binding
agreements. The game structure is as Figure 5.1 shows.

147

Chapter 6

6.3.2 The mathematical model

6.3.2.1 The retailers model
We first consider the payoff function Z rl (l=1, 2,L) for the retailers. The
retailers objective is to maximize his net profit by optimizing his strategy xrl ,
including replenishment decision and profit margin. The relevant parameters and
variables of the retailer l are designed in Table 6.1.
Table 6.1. Parameters and variables for retailers

Drl : Retailer ls annual demand

Arl : A constant in the demand function of retailer l, which represents his market scale
erl : Coefficient of the products demand elasticity for retailer l
Orl : Ordering processing cost for retailer l per order of product l
Rrl : Retailer ls annual fixed costs for the facilities and organization to carry this product
prl : Retail price charged to the customer by retailer l
krl : Decision valuable, the integer divisor used to determine the replenishment cycle of retailer l
Grl : Decision valuable, retailer ls profit margin

The retailer l faces the holding cost, the ordering cost and an annual fixed cost. The
inventory level for retailer l is shown as Figure 6.2(a). The retailer ls objective function
is identical with Chapter 5 given as:

m a xZ r Gr Dr
krl ,Grl

TDrl
2krl

hr l

Or lkr
T

Rr

(6.1)

Subject to

krl {1, 2,3,...} ,

(6.2)

Grl prl pml ,

(6.3)

Drl Arl erl prl ,

(6.4)

Grl 0 ,

(6.5)

148

Chapter 6

0 Drl Pl .

(6.6)

Constraint (6.2) gives the value of the divisor used to determine the retailer ls
replenishment cycle time. Constraint (6.3) indicates the relationship between the prices
(the retail price prl and the wholesale price pml ) and retailer ls profit margin.
Constraint (6.4) is the demand function at retailer l which is almost invariably a
downward sloping and convex function with respect to the retail price in real world.
Constraint (6.5) ensures that the value of Grl is nonnegative. Constraint (6.6) gives the
bounds of the annual demand, which cannot exceed the annual production capacity Pl
of the product.
Inventory level

TDrl
krl
Order point

T
krl

Time

(a) Retailer ls inventory level of final product

Inventory level

TDrl
Pl

T
krl

(b) Manufacturers inventory level for product l

Figure 6.2. Inventory levels of retailer and manufacturer

149

Time

Chapter 6

6.3.2.2 The manufacturers model

The manufacturers objective is to determine his strategy xm , composed of the
setup time interval for production, the profit margins for all the products and the
selection decision of suppliers and components, to maximize his net profit. The
relevant parameters and decisions variables are shown in Table 6.2.
Table 6.2. Parameters and variables for manufacturer

Pl : Annual production capacity for product l

cml : Production cost per unit product l
hmpl : Holding costs per unit of product l inventory

S m : Setup cost per production

FLij : Fixed cost of using component Lij
uLik l : Usage amount of unit component Lij per unit product l
wsv : The fixed cost of using supplier v, covering supplier certification, contract setup, etc.

Rm : Annual fixed costs for the facilities and organization for the production of the products
pml : Wholesale price charged to retailer l
T : Decision variable, manufacturers setup time interval
Gml : Decision variable, manufacturers profit margin for product l

ijk : Binary decision variable to indicate whether component Lij has been used to replace Lik

s : Binary decision variable to indicate whether supplier v is used

v

tsv Lij : Binary decision variable to indicate whether component Lij is supplied by supplier v
z Lij : Binary decision variable to indicate whether component Lij is used

The manufacturer faces the costs of components used, production cost, fixed costs
to contract with the suppliers and using components, holding costs for the products,
setup cost, and an annual fixed cost. According to the given VMI policy, the
manufacturer does not have to pay for the inventory cost for components. The behavior
of the inventory level for the product for the manufacturer is illustrated as Figure 6.2(b).
The annual inventory for product ls is given by

150

T
Dr
2 l

1 Dr
1 l (Lu 1995). The
kr
Pl
l

Chapter 6

setup cost S m occurs at the beginning of each production. Thus, we can easily derive
the manufacturers payoff function Z m :

max

Gml ,T , sv , tsv Lij , zLij , ijk

l 1

v 1

Ni

Z m Gml Drl wsv sv FLij zLij

i 1 j 1

T L
Dr
2 l 1 l

1 Dr
1 l hmpl
kr
Pl
l

S
m Rm
T

(6.7)
Subject to
Ni
I Ni 1

I
Gml pml pLij zLij uLij l uLik l ijk pLiN zLiN uLiN l cml ,
i
i
i
i 1 j 1
k j 1

i 1
k 1

j 1

ijk

z Lik 1 , uLik l 0 ; i 1, 2,..., I ; j 1, 2,..., Ni 1; k j 1,..., Ni

(6.8)

(6.9)

ijk zL , i 1, 2,..., I ; j 1, 2,..., Ni 1; k j 1,..., Ni

(6.10)

tsv Lij 0 , Lij Qsv , v 1, 2,...,V

(6.11)

ij

sv Lij

tsv Lij sv sv , v 1, 2,...,V

v 1

Lij Qsv

zLij , i 1, 2,..., I ; j 1, 2,..., Ni

(6.12)
(6.13)

zLij , tsv Lij , sv , ijk 0,1 , i 1, 2,..., I ; j 1, 2,..., Ni 1; k j 1,..., Ni

(6.14)

Gml 0 , T 0 .

(6.15)

l 1, 2,..., L

Constraint (6.8) gives the relationship between the prices (the wholesale price and
the component prices) and the manufacturers profit margin. As indicated in the fourth
point of the assumption in Section 6.2, Constraint (6.9) ensures that if a component is
the lowest functionality components for any product, it is either used or replaced, but
not both. Constraint (6.10) makes sure that only procured components can be used to
replace other components. (6.9) and (6.10) together ensure that the demands for all

151

Chapter 6

components are satisfied. Also, they met the one-way substitutability constraint which
ensures that a higher functionality component can replace a lower functionality
component but not vice versa. Constraint (6.11) sets the value of tsv Lij 0 for all
components Lij Qsv for all the suppliers. Constraint (6.12) indicates that a component
is procured from exactly one supplier. Constraint (6.13) sets the value of sv to 1, if
supplier v supplies a component, and ensures that the number of different types of
components supplied by supplier v is no greater than sv . The value ranges of all the
variables are set by constraints (6.14) and (6.15).
6.3.2.3 The suppliers model
Each suppliers problem is to determine an optimal strategy xsv ( v 1, 2,...,V ),
including pricing decisions for the components supplied, to maximize his net profit.
Table 6.3 gives the parameters and decision variables for the suppliers.
Table 6.3. Parameters and variables for suppliers

sv : Number of different types of components supplier v is capable of supplying

cLij : The cost of component Lij paid by supplier

Qsv : Set of components supplied by supplier v

Rsv : Supplier vs annual fixed costs for the facilities and organization to carry the components

pLij : Decision variable, the price of component Lij charged by the supplier to the manufacturer
The supplier faces component costs and an annual fixed cost. According to the
given VMI policy, the suppliers do not pay for inventory cost. Therefore, the supplier
vs payoff function Z sv is:
L

Ni 1

max Z s pL cL zL uL l
p , L Q
Lij

ij

sv

l 1 i 1

j 1

ij

ij

ij

ij

Ni

k j 1

Dr ts L pL cL

Lik l ijk

v ij

iNi

iNi

LiNi

uLiN l Drl tsv LiN Rsv sv

i
i

(6.16)

152

Chapter 6

Subject to

pLij 0 ,

Lij Qsv

(6.17)

Constraint (6.17) ensures the non-negativeness of pLij . Here, we do not constraint

pLij to be larger than its cost. That is because, the supplier may lower the price for one
component (even lower than its cost) to attract the manufacturer to buy his other
components.
6.3.2.4 Game model
Based on the game structure shown in Figure 5.1, we formulate the three-level
nested Nash game model as follows:

M ax Z r , for each l 1, 2,..., L ,

xrl

Subject to:
Constraints (6.2)-(6.6);

M ax Z m ,
xm

Subject to:
Constraints (6.8)-(6.15);

M ax Z s , for each v 1, 2,...,V ,

xsv

Subject to:
Constraint (6.17).
In this case, the retailers and suppliers have the same net profit structures, so we can
use one uniform formula to represent the payoff function Z rl of each retailer as
defined in (6.1)-(6.6) and one uniform formula for each suppliers payoff function

Z sv as in (6.16)-(6.17). In the above game model, payoff function Z rl and constraints

(6.2)-(6.6) for each retailer l 1, 2,..., L in this game model formulate the RMS game.
Payoff function Z m and constraints (6.8)-(6.15) formulate the MS game, which is the

153

Chapter 6

model of the manufacturer. The SS game is formulated by payoff function Z sv and

constraint (6.17) for each supplier v 1, 2,...,V .

6.4 SOLUTION ALGORITHM

6.4.1 Reaction functions

6.4.1.1 The retailers reactions
Retailer ls demand function can be expressed by the corresponding profit margins.
Substituting (6.2) and (6.8), we can rewrite (6.4) as:
Ni
I Ni 1

I
Drl Arl erl Grl Gml pLij zLij uLijl uLik l ijk pLiN zLiN uLiN l cml
i
i
i

i 1 j 1
k j 1

i 1

. (6.18)

Now suppose that the decision variables for the suppliers and the manufacturer are
fixed. According to Viswanathan and Wang (2003), the best reaction krl that
maximizes the retailers net profit Z rl can be expressed as:

2T 2 hrl Drl
krl * 1 1
Orl

/ 2 .

(6.19)

Here, a is the largest integer no larger than a.

We then consider the optimal value of Grl . Substituted (6.18) into (6.2) and set the
first derivative of Z rl with respect to Grl equal to zero. We have:

Grl

Cl hrl T

,
2erl 4krl

(6.20)

Ni
I Ni 1

I
where Cl Arl erl Gml pLij zLij uLijl uLik lijk pLiN zLiN uLiN l cml
i
i
i

i 1 j 1
k j 1

i 1

Since the second derivative of Z rl with respect to Grl is negative, Z rl is a

154

Chapter 6

concave function of Grl . From constraints (6.5) and (6.6), we can obtain lower bound
and the upper bound of Grl :
Ni
I Ni 1

Ar Pl

I
Grl max 0, l
Gml pLij zLij uLij l uLik lijk pLiN zLiN uLiN l cml
i
i
i

erl
i 1 j 1
k j 1

i 1

(6.21)
Ni
I Ni 1

I
Grl
Gml pLij zLij uLij l uLik lijk pLiN zLiN uLiN l cml
i
i
i
erl
i 1 j 1
k j 1

i 1

Arl

(6.22)

If Grl obtained from (6.20) is in the interval of Grl , Grl , it is obviously the

optimal reaction Grl * of the retailer. Otherwise, we have to substitute the bounds (6.21)
and (6.22) into (6.2), the bound that provides higher profit is the best reaction Grl * .
6.4.1.2 The manufacturers reactions
Assume that the decision variables for the suppliers and the retailers are fixed.
Substituting (6.18) into (6.7), taking the first derivative of Z m with respect to T and

Gml and setting the result to zero, one has:

T

2Sm

Drl

l 1

1 Dr
1 l hmpl
krl Pl

hmpl T
1
1
2 krl
Gml
hmp T
1 2er
2 l 1 l
2 krl
Pl

(6.23)

Wml

where Wml

Since

Wml ,

Ni
I Ni 1

I
Grl pLij zLij uLij l uLik lijk pLiN zLiN uLiN l cml
i
i
i
erl
i 1 j 1
k j 1

i 1

Arl

(6.24)

2 Zm
2S
3m 0 , Z m is a concave function of T . Thus, equation (6.23)
2
T
T

155

Chapter 6

provides the optimal T * .

Z m s second derivative with Gml is

2 Zm
1
2erl 1
2
kr
Gml
l

erl 2 hmpl T
.
erl T

P
l

(6.25)

2Zm
If
0 , the optimal Gml can be obtained (6.24):
Gml 2

From constraints (6.6) and (6.15), we can obtain lower and the upper bounds of

Gml :
Ni
I Ni 1

Ar Pl

I
Gml max 0, l
Grl pLij zLij uLij l uLik lijk pLiN zLiN uLiN l cml
i
i
i

erl
i 1 j 1
k j 1

i 1

(6.26)
Ni
I Ni 1

I
Gml
Grl pLij zLij uLij l uLik lijk pLiN zLiN uLiN l cml
i
i
i
erl
i 1 j 1
k j 1

i 1

Arl

. (6.27)

If Gml obtained from (6.24) is in the interval of Gml , Gml , it is the optimal

reaction Gml * of the manufacturer. Otherwise, Z m reaches its maximal value when

Gml is at its upper bound or lower bound. The bound that provides higher profit is the
optimal reaction Gml * .
If

2Zm
0 , we only have to find the bound that provides maximal value of Z m .
Gml 2

That is the best reaction Gml * .

The manufacturers problem to find optimal suppliers and components to maximize
his profit is 0-1 integer programming with objective function (6.7) and Constraints
(6.9)-(6.14). Some optimization software, such as Lingo, etc., or some search algorithm,
can be used to solve this problem to obtain the best reactions ijk * , sv * , tsv Lij * , and
156

Chapter 6

z Lij * .
6.4.1.3 The suppliers reactions
Lastly, we consider the reaction functions for the suppliers. Suppose that the
decision variables for the retailers and the manufacturer are fixed. The suppliers
problem is to find the optimal reaction for prices for the components his supplies. The
second order condition for Z sv in (6.16) is negative, thus, the necessary condition to
maximize the suppliers net profit Z sv is:
Nm
L

z
u

Lmn Lmnl uLmk l mnk Drl tsv Lmn erl l , 1 n N m 1

l 1
k n 1

zLmN uLmN l Drl tsv LmN erl l ,

n Nm

m
m
m
l 1
0

Z sv
pLmn

(6.28)

Ni
I Ni 1

where l pLij cLij tsv Lij zLij uLijl uLik lijk pLiN cLiN tsv LiN zLiN uLiN l .
i
i
i
i
i

i 1 j 1
k j 1

Substitute (6.18) into (6.28), we can obtain:

Nm
L

uLmnl uLmk l mnk BLmnl tsv Lmn erl Lmnl

k n 1

l 1
,1 n N m 1
2
Nm

2er z L uL l u L l mnk t s L
l
mn
mn
mk
v mn

l 1
k n 1

uLmN l BLmN l tsv LmN erl LmN l

m
m
m
m
l

,
n Nm
L

2
2erl z LmN uLmN l tsv LmN

m
m
m
l 1

pLmn

where

BLmnl

Ni
I Ni 1

Grl Gml pLij zLij uLijl uLik lijk

i 1 j 1
k j 1

im

Arl erl
Nm
Nm 1

I
pLmj zLmj uLmj l uLmk l mjk pLiNi z LiNi uLiNi l cml
k j 1
jj 1n

i 1

157

(6.29)

Chapter 6

Ni
Nm
I Ni 1

Nm 1

Lmnl pLij cLij tsv Lij zLij uLijl uLik l ijk pLmj cLmj tsv Lmj zLmj uLmjl uLmk l mjk
i 1 j 1
k j 1
k j 1

jj 1n

im

Nm

I
tsv Lmn zLmn cLmn uLmnl uLmk l mnk pLiN cLiN tsv LiN zLiN uLiN l
i
i
i
i
i
k n 1

i1

BLmN

Ni
I Ni 1

I
Arl erl Grl Gml pLij zLij uLij l uLik lijk pLiN zLiN uLiN l cml
i
i
i

i 1 j 1
k j 1

ii 1m

Ni
I Ni 1

LmN l pLij cLij tsv Lij z Lij uLij l uLik lijk

m
i 1 j 1
k j 1

pLiN cLiN tsv LiN z LiN uLiN l cLmN tsv LmN z LmN uLmN l .
i 1
im

From constraints (6.6) and (6.17), we can obtain the upper bound and lower
bound of pLmn :

pLmn

BLmnl Pl
max 0, max
,1 n N 1
m
Nm

erl z Lmn uL l uLmk l mnk

mn
k n 1

BLmN l Pl
m
max 0, max
,
n Nm
l

erl z LmN uLmN l

m
m

(6.30)

pLmn

BLmnl

,1 n N m 1
min
Nm
l

erl z Lmn uL l uLmk l mnk

mn
k n 1

BLmN l
m
min
,
n Nm
l
erl z LmN uLmN l

m
m

(6.31)

If pLmn obtained from (6.30) is in the interval of pLmn , pLmn , it is the best

reaction pLmn * of supplier v. Otherwise, we have to substitute the bounds (6.30) and
(6.31) into (6.11), the bound that provides higher profit is the optimal reaction pLmn * .

158

Chapter 6

6.4.2 Algorithm for the Nash equilibrium of RMS game

As section 6.3.1, we denote X rl , X m and X sv as the strategy sets of retailer l,
manufacturer and supplier v, respectively. X s X s1 ... X sV , X r X r1 ... X rL ,
X ms X s X m and X X ms X r are the strategy profile sets of the suppliers, the

retailers, the suppliers and the manufacturer, and all the chain members. The following
procedure is used for solving the three-level nested Nash game model:
Step 0. Initialize x(0)

(0)

, xm(0) , xr (0)

in the strategy profile set X .

Step 1. Denote xr l (0) as the strategy profile of all the chain members in x (0) except
for retailer l. For each retailer l, fixed xr l (0) , using xrl (0) , find the optimal reaction

xrl * krl * , Grl *

(as Section 6.4.1.1) to optimize the retailer ls payoff function Z rl in

its strategy set X rl .

Step 2. Denote x m (0) as the strategy profile of all the chain members in x (0) except
for the manufacturer. Fixed

xm* T * , Gml * , ijk * , sv * , tsv Lij * , zLij *

x m (0) , using

xm (0) , find the optimal reaction

(as Section 6.4.1.2) to optimize the profit function

Z m in its strategy set X m .

Step 3. Denote xs v (0) as the strategy profile of all the chain members in x (0) except
for supplier v. For each supplier v, fixed xs v (0) , using xsv (0) , find the optimal reaction

xsv * pLmn * , for all Lmn Qsv (as Section 6.4.1.3), to optimize the profit function

Z sv in its strategy set X sv . If xs* xs (0) 1 , the Nash equilibrium of SS game, xs* ,
obtained, Go step 4. Otherwise, xs (0) xs* , repeat step 3.
Step 4. xms* xs* , xm* . If

xms* xms (0) 2 , the Nash equilibrium of MS game,

159

Chapter 6

xms* , obtained, Go step 5. Otherwise, xms (0) xms* , go step 2.

Step 5. x* xms* , xr * . If

x* x(0) 3 , the Nash equilibrium of RMS game, x* ,

obtained. Output the optimal results and stop. Otherwise, x(0) x* , go step 1.
(: 1 , 2 , 3 are very small positive numbers)

6.5 NUMERICAL RESULTS

To demonstrate the application of the proposed game model and solution algorithm,
we present and discuss a specific case in a computer industry. This particular case is
concerned with coordinating suppliers and components selection, pricing and
replenishment decisions of the supply chain (as Figure 6.1) described in Section 6.2.
Through this example, we are also interested in investigating the impacts of market
scale parameter and components selection on the decisions and profits of the suppliers,
the manufacturer and the retailers.
6.5.1 Initialization
The explicit numerical parameters selected for the base case example reflect those
shown in Lu 1995, Woo et al. 2001, Grave and Willems 2005, Prafulla et al. 2006. For
the two retailers in European market and North American market (i.e., r1 , r2 ), we set:

Ar1 Ar2 2 106 ; er1 1400 , er2 1200 ; Or1 4 , Or2 3 ; hr1 1 , hr2 2 . For the
manufacturer, (a) setup cost and ordering cost are: Sm 95 , Om 5 ; (b) holding costs
for the two products are: hmp1 0.5 , hmp2 1 ;

(c) production capacity are:

P1 5 106 , P2 3 106 ; (d) production costs are: cm1 10 , cm2 5 ; (e) fixed costs of

160

Chapter 6

using suppliers are set at: ws1 12 , ws2 13 , ws3 40 , ws4 60 , ws5 18 , ws6 31 ;
and (f) fixed costs of using components are set at: FL11 3 , FL12 4 , FL21 6 ,

FL22 7 , FL23 8 , FL31 9 , FL41 12 , FL51 2 , FL61 10 . The suppliers purchase

costs for the components are: cL11 100 , cL12 50 , cL21 200 , cL22 80 , cL23 50 ,

cL31 20 , cL41 30 , cL51 200 , cL61 29 . The fixed costs for all the suppliers,
manufacturer and retailers are: Rs1 30 , Rs2 15 , Rs3 35 , Rs4 23 , Rs5 20 ,

Rs6 18 , Rm 100 , Rr1 50 , Rr2 61 . The relationship between the components and
their suppliers are shown as Figure 6.1.
By applying the above solution procedure in section 6.4.2, the optimal results for
the suppliers, the manufacturer and the retailers in this example are given as Table 6.4
and Table 6.5. In order to ensure that our conclusions are not based purely on the
chosen numerical values of the base example, we conduct sensitivity analysis on market
scale parameter, setup cost and component cost. In addition, we also identify the
influence of component selection strategy. Two cases are analyzed, no components
selection case and with components selection case. For the first one, components used
for each product have already been fixed a priori. No substitution could be conducted.
For the second one, although components used for each product have been fixed, higher
functionality components can be used to substitute lower ones. We denote them as NCS
case and WCS case, respectively.

161

Chapter 6

6.5.2 Simulation Results

6.5.2.1 Sensitivity analysis for market scale parameter
Let us first consider the market scale parameter Ar1 . It can be seen from Table 6.4,
when Ar1 is 2.0 106 , the same suppliers and components are selected for the NCS
case and WCS case. As Ar1 increases, less suppliers are selected and higher
functionality components are used to replace lower ones in the WCS case. For example,
when Ar1 increases to 2.5 106 , component L23 is substituted by L22 for product 2.
Moreover, when Ar1 is 5.0 106 , the highest functionality components L11 and L21
are used to substitute lower functionality components L12 and L22 for both products.
In this case, supplier 2 and supplier 3 are not selected, while supplier 4 is selected.
(b) Wholesale price for Product 2

(a) Wholesale price for Product 1

1800

1200
1,059.70

Price

Price

1,572.60

1,140.10 1,418.80
843.9

782.92
782.92

1,105.90

791.03
789.68

823.57
803.17

839.31

695.83
836.24

NCS
WCS

695.83

NCS
WCS

Ar1

400
2

2.5

3.5

(106)

Ar1

400
2

2.5

3.5

(106)

Figure 6.3. Influence of market scale parameter on product wholesale prices

From Figure 6.3, we can see that the wholesale prices for both products will
increase as Ar1 increases and product 1s price increases more significantly. For
example, in the NCS case, when Ar1 increases to 3.5 106 from 2.0 106 , the
wholesale price for product 1 increases by 63.85%, while for product 2, the price only
raise by 2.59%. Meanwhile the wholesale price for product 1 in NCS case is higher
162

Chapter 6

than WCS case, whereas for product 2, the result reverses. The variation trend is similar
for the retail price. That is to say the component selection strategy reduces the prices
for product 1 and raises the product 2s prices.
As Figure 6.4 shows, the suppliers total profit and retailer 2s profit in the NCS
case are higher than the WCS case. The manufacturers profit and retailer 1s profit are
higher in WCS case when Ar1 increases. Moreover, we can see that for the above
profits, their profit disparity between the NCS case and the WCS case increases with
the increase of Ar1 . For example, when Ar1 is 2.5 106 , the suppliers total profit in
the NCS case is only 7.27% greater than that in the WCS case; the manufacturers
profit in the NCS case is 6.00% less than WCS case. However, when Ar1 increases to
5.0 106 , the profit gap is extended to 64.83% and 66.95%, respectively. Thus, the

component selection strategy increases the manufacturers and retailer 1s profits while
compressing the suppliers total profit and retailer 2s profit.
From Table 6.5, we can see that the increase of Ar1 shortens the setup time interval
for the manufacturer and the replenishment cycle for retailer 1 in NCS case, while it is
not necessary the same for the manufacturer and retailer in WCS case. For instance, in
NCS case, when Ar1 increases from 3.5 106 to 5.0 106 , the cycles for the
manufacturer and retailer 1 reduce from 0.0356 to 0.0332 and from 0.0356/12 to
0.0332/14, respectively. However, in WCS case, the manufacturers replenishment
cycle increases from 0.0358 to 0.1195.

163

Chapter 6

(a) Total profit for suppliers

(107)

200

80
NCS
WCS

(b) Profit for manufacturer

(107)

NCS
WCS

149.47

Profit

Profit

53.36

73.882
90.685
28.0286
28.0286

39.8667

16.826

67.3796
5.6451

37.1659

Ar1

2.5

3.5

(106)

(c) Profit for retailer 1

(107)

5.6451

NCS
WCS

7.6617
2.5

Profit

162.18

23.43
23.43

23.073
23.002

64.731
18.791
0

18.791
2

31.045

Ar1

3.5

22.369

20.536

11.053

Ar1

3.5

(106)

21.324

68.021

31.552
2.5

14.342

NCS
WCS

139.83

Profit

8.1211

(d) Profit for retailer 2

(107)

40

200

31.961

(106)

0
2

2.5

3.5

(106)

Ar1

Figure 6.4. Influence of market scale parameter on profits for chain members

6.5.2.2 Sensitivity analysis for production setup cost

Let us now consider the production parameter the manufacturers setup cost S m .
Table 6.6 and 6.7 summarize the optimal decisions and profits for NCS case for all the
supply chain members, when S m increases to 200, 300 and 400 from 100 in base
example. There is no component substitution occurs. The variation of the optimal
decisions and profits are almost the same for both NCS and WCS cases. Thus, we only
discuss the results for NCS case.
From Table 6.6, we can see that product 1s wholesale price has a very slightly

164

Chapter 6

increase, while for product 2, the price decreases. There is no significant change for
retail price and profit for product 1. The retail price for product 2 decreases and retailer
2s profit increases slightly. The larger demand from product 2 does not increase
profitability for the manufacturer. For example, when S m increases from 200 to 300,
the wholesale price for product 1 increases from 695.8392 to 695.8426 and for product
2, the wholesale price reduces from 782.5831 to 782.3564. The manufacturers profit
still reduces from 5.5710107 to 5.5151107.
It can be seen from Table 6.6, all the suppliers profits increase except for supplier 1.
The reason is that the component provided by supplier 1 is only used for product 1,
whose demand has a slightly decrease.
The setup time interval for the manufacturer has significantly lengthen as the
increase of S m . For example, when S m increases from 200 to 300, the setup time
interval increases from 0.0579 to 0.0708, 18.22%.
6.5.2.3 Sensitivity analysis for component cost
To investigate the influence of component cost on the optimal decisions and profits
for the supply chain members, without loss of generality, we conduct sensitivity
analysis by increasing the component L12 s cost by 50% at a time. This component is
selected for product 2 as a priori and could also be substituted by higher functionality
component L11 for product 2. All the results for component L12 s cost at 75, 100 and
125 are given as Table 6.8 and 6.9.
Figure 6.5 presents the optimal wholesale prices for product 1 and product 2. We

165

Chapter 6

can see that as the cost for component L12 increases, the product 1s wholesale price
decreases and product 2s wholesale price increases in NCS case. However, the
wholesale price for product 2 in WCS case might reduce. For example, when the
component cost increases from 50 to 75, product 2s wholesale price decreases from
782.92 to 766.76. This is due to the substitution of component L11 to L12 in WCS
case, which reduces the cost for product 2 to the manufacturer and brings larger
demand for product 2. The same results can be derived for retail prices. We do not
describe in figures.
(b) Wholesale price for product 2

(a) Wholesale price for product 1

700

810
798.73

Price

695.83

695.16
694.87

695.16
694.22

793.45

695.17

789.22
782.92

Price

695.83

693.41

782.92
766.76

766.76

NCS
WCS
685
50

75

100

125

766.78

NCS
WCS

cL12

750
50

75

100

125

cL12

Figure 6.5. Influence of component cost on product wholesale prices

We then consider the variation of chain members profits under the change of
component L11 s cost. From Figure 6.6, it can be seen that the retailer 1s profit in
WCS case is lower than NCS case, while for retailer 2, it is reversed. For the suppliers
total profit and the manufacturers profit, in this situation, only if the component L11 s
cost is larger than some value, the suppliers total profit and manufacturers profit in
WCS case could be larger than those in NCS case. In our example here, this value is
between 100 and 125.
166

Chapter 6

From Table 6.9, we can easily see that the increase of component cost will lengthen
the setup time interval for the manufacturer. Besides, the setup time interval for the
manufacturer is shorter in the WCS case. This is due to the larger demand from product
2 in WCS case.
30

(a) Total profit for suppliers

(107)

6
NCS
WCS

28.0286

5.6451
5.6451

Profit

Profit

28.0286

(b) Profit for manufacturer

(107)

26.9132

5.1376
5.05

26.1764
25.3957

19

4.8898
NCS
WCS

100

125

cL12

50

(c) Profit for retailer 1

(107)

5.0501

25.2636

24
75

5.0512

25.3957
25.3958

50

5.3383

25

75

100

125

cL12

(d) Profit for retailer 2

(107)

18.916

23.43

Profit

18.841

18.874

Profit

18.791
18.791

18.826

18.826

18.825

23.43

24.295
23.097

24.295

22.875

24.294

22.599
NCS
WCS

NCS
WCS
18.5
50

75

100

125

20

cL12

50

75

100

125

cL12

Figure 6.6. Influence of component cost on profits for chain members

6.5.3 Managerial implications

Based on the results obtained from the simulation results presented above, in this
subsection, we wrap up our findings and some main implications.
Firstly, under WCS case, the increase of one retailers market scale leads the
manufacturer to use higher functionality components for the product of this market.

167

Chapter 6

This implies that when one retailers market scale is large, the manufacturer would like
to provide this retailer with the product that is made of high functionality components
substituting low functionality ones. Otherwise, the manufacturer would build up the
product exactly the same as this retailer requires.
Secondly, when Arl is large, the manufacturer could reap the benefit from the
components selection strategy, while it is opposite for the whole supplier sector, and the
supply chain system: they may suffer loss. Compared with NCS case, when Arl is
large, the manufacturer tends to select less components with higher functionality to
achieve lower wholesale price for the product of this market and obtain higher profit
from this product, which compresses the total profits of the suppliers. The increased
profits
for the manufacturer and the retailer could not cover up the loss for the suppliers.
Hence, the system profit also decreases.
Thirdly, when Arl increases, the manufacturer and the retailer could obtain bigger
profit in WCS case. In the contrary, the suppliers as a group, the other retailer and the
supply chain system benefit more from NCS case. That is because the sum of the
profits of the suppliers and the system profit increase more significantly in NCS case
and the profits for the manufacturer and the retailer increase more significantly in WCS
case. And the other retailers profit in WCS case decreases more significantly than the
NCS case.
Fourthly, the increase of production setup cost S m may benefit suppliers. This is in

168

Chapter 6

accordance with the observation in Chapter 5. The reason is that as S m increase, the
manufacturer would like to reduce the wholesale prices for some products in order to
simulate the market demands. Thus, the larger demand for components brings more
profits for some suppliers.
Fifthly, the increase of Arl might lengthen the setup time interval for the
manufacturer in WCE case. This is because the manufacturer will use higher
functionality component to replace lower ones, which might reduce the market demand
for the higher price. Thus, the manufacturer would setup production less frequently.
Lastly, when component cost increases, the manufacturer would produce the
products more frequently in WCE case than NCS case. This is simplicity due to the fact
that the manufacturer could obtain component at a lower price through substitution in
WCE case.

6.6 SUMMARY

In this chapter, we have considered the coordination of supplier selection, pricing,

and replenishment decisions in a multi-level supply chain composed of multiple
suppliers, one single manufacturer and multiple retailers. This coordination problem is
modeled as a three-level nested Nash game model. We use both analytical and
computational methods for the derivative of the optimal decisions of all the chain
members. A numerical study and sensitivity analysis are conducted to examine the
game model and solution algorithm. Findings observed from computational sensitivity

169

Chapter 6

and theoretical analyses generate a series of managerial implications: (a) when one
retailers market scale increases, the manufacturer tends to use high-end components
for the product sold to this market. As a result, the manufacturer would benefit from
this strategy and the supply chain system includes a fewer number of suppliers and their
total profit becomes lower; (b) the increase of manufacturers production setup cost
might benefit the suppliers; (c) the manufacturer would setup his production more
frequently when component selection strategy is employed.

170

Table 6.4. Results for suppliers under different market scale

(a) NCS case
s1

Supplier & Supplied

components

Arl

L11

Results

Base example
6

2.510

3.510

5.010

s2
L21

L12

s3
L23

L61

s4

L22

L41

L31

Price
Profit (107)
Price
Profit (107)
Price
Profit (107)

138.4655
1.9729
168.9760
4.5474

137.9322

88.4629
6.6354
131.1628 118.9718
8.8178

129.4662 68.4595
6.6349
92.2132
98.9576
8.8173

229.9612
12.372

117.6704

17.7369 159.9449
15.877

Price
Profit (107)

333.0985
32.614

81.5407

179.9597
15.878
283.0949
34.179

s5
L41

L51

L31
85.5214

15.7579
6.3924

114.5853

261.2852
6.3930
274.5903
8.8424

4.589510-13
8.8418

141.1994

301.2043
14.878

7.779710-13
14.877

180.0559

340.0620
26.550

3.973810-13
26.549

206.9897
29.578

s6
L61

171

(b) WCS case

Supplier & Supplied
components

s1

Arl

Results
Base example
Price
Profit (107)
6
Price
2.510
Profit (107)
6

3.510

Price
7

L21

L12

138.4655
1.9729
176.6294
5.0930

137.9322

264.0987

97.2780

s3
L23

88.4629
6.6354

129.8218
4.1937

L61

s4

L22

L41

129.4662
68.4595
6.6349
93.6614
94.4361
9.2937
71.8073

L41
0
0

148.3697

Profit (10 )
Price

16.014
207.9108 213.2585

2.3916

16.325

Profit (107)

22.671

: implies the component is not selected for the final products

L31

s5

0
151.1665
22.671

L51

s6
L31

L61

261.2852
6.3930
278.0925
9.2931

85.5214 15.7579
6.3924
75.1250 42.9616
9.2924

310.1731

52.6335

16.325
321.1705

16.324
75.1250 83.3174

22.671

22.672

97.5349
Chapter 6

5.010

L11

s2

Table 6.5. Results for manufacturer and retailers under different market scale
Manufacturer &
Retailer

NCS case
m

Arl

Results

Product 1

Base
example

Price
Cycle
Demand(105)
Profit(107)
Price

695.8341
782.9200
0.0411
0.0411
5.1291
5.3025
5.6451

2.510

Cycle
5

Demand(10 )
172

3.510

5.010

Profit(10 )
Price
Cycle
Demand(105)
Profit(107)
Price
Cycle
Demand(105)
Profit(107)

Product 2

r1

r2

1,062.2
0.0411/10
5.1291
18.791

1,224.8
0.0411/17
5.3025
23.430

Product 1

Product 2

695.8341
782.9200
0.0411
0.0411
5.1291
5.3025
5.6451

r1

r2

1,062.2
0.0411/10
5.1291
18.791

1,224.8
0.0411/17
5.3025
23.430

843.8991

789.6813

1,314.8

1,228.2

836.2433

791.0269

1,311.0

1,228.8

0.0389

0.0389

0.0389/11

0.0389/16

0.0384

0.0384

0.0384/11

0.0384/16

6.5927

5.2619

6.5927

5.2619

6.6463

5.2538

6.6463

5.2538

31.045

23.073

31.552

23.002

1,803.0
0.0358/13
9.7586
68.021

1,245.1
0.0358/15
5.0586
21.324

2,495.1
0.1195/52
15.068
162.18

1,363.2
0.1195/42
3.6419
11.053

WCS case

7.6617

8.1211

1,140.1
0.0356
9.5196

803.1702
0.0356
5.1810
14.342

1,820.0
0.0356/12
9.5196
64.731

1,234.9
0.0356/15
5.1810
22.369

1,105.9
0.0358
9.7586

1,572.6
0.0332
13.992

839.3057
0.0332
4.9642
31.961

2,572.0
0.0332/14
13.992
139.83

1,253.0
0.0332/14
4.9642
20.536

1,418.8
0.1195
15.068

823.5675
0.0358
5.0586
16.826
1,059.7
0.1195
3.6419
53.360

Chapter 6

Table 6.6. Results for suppliers under different production setup cost
s1

Supplier & Supplied

components

Sm
200
300
400

L11

Results
Price
Profit (107)
Price
Profit (107)
Price
Profit (107)

s2
L21

138.4713
1.9732
138.4676
1.9730
138.4652
1.9729

L12

s3
L23

138.2184

88.4655
6.6525
138.4452 88.4618
6.6655
138.6362 88.4622
6.6767

L61

L22

s4
L41

129.7783 68.4473
6.6519
130.0079 68.4441
6.6650
130.2080 68.4409
6.6762

L31

s5
L41

0
0
0

s6

L51

L31

261.4155
6.4078
261.5183
6.4194
261.6062
6.4293

85.6242

L61

15.6590
6.4072
85.7042
15.6424
6.4187
85.8017
15.6389
6.4284

Table 6.7. Results for manufacturer and retailers under different production setup cost
Manufacturer &
Retailer
173

Sm
200

Results
Price
Cycle
5

Demand(10 )
7

300

r1

r2

782.5831

1,062.2

1,224.6

0.0579

0.0579

0.0579/15

0.0579/24

5.1291

5.3045

5.1291

5.3045

18.791
1,062.2
0.0708/18
5.1291
18.791
1,062.2
0.0817/21
5.1291
18.791

23.448
1,224.5
0.0708/30
5.3058
23.460
1,224.4
0.0817/34
5.3070
23.470

Product 1

Product 2

695.8392

5.5710
695.8426
782.3564
0.0708
0.0708
5.1291
5.3058
5.5151
695.8469
782.1654
0.0817
0.0817
5.1291
5.3070
5.4676

Chapter 6

400

Profit(10 )
Price
Cycle
Demand(105)
Profit(107)
Price
Cycle
Demand(105)
Profit(107)

Table 6.8. Results for suppliers under different component cost

(a) NCS case
s1

Supplier & Supplied

components

cL12
75
100
125

Results
Price
Profit (107)
Price
Profit (107)
Price
Profit (107)

L11

s2
L21

139.4455
2.0259
140.0923
2.0609
140.9013
2.1048

L12

L23

s3
L61

161.5828

89.4394
6.3207
177.3546 90.0862
6.1135
197.0682 90.8952
5.8575

L22

s4
L41

L31

122.1698
69.4208
6.3201
117.2948
70.0676
6.1129
111.1994
70.8765
5.8570

s5
L41

L51

s6
L31

L61

258.8774
6.1236
257.2741
5.9449
255.2697
5.7225

83.1167
15.7541
6.1229
81.5145
15.7530
5.9442
79.5105
15.7527
5.7218

s4

s5

s6

0
0

(b) WCS case

s1

174

Supplier & Supplied

components

cL12
75
100
125

Results

L11

s2
L21

L12

L23

s3
L61

L22

L41

Price
Profit (107)

155.2926
5.8241

86.6337
1.8807

120.4483
66.6278
6.0421

Price
Profit (107)
Price
Profit (107)

155.2928
5.8242
155.3076
5.8256

86.6314
1.8806
86.6496
1.8815

120.4420
66.6307
6.0421
120.4816
66.6161
6.0425

L31

L41
0
0
0

L51

L31

L61

255.2962
5.8245

77.5107 17.7835
5.8243

255.2965
5.8246
255.2900
5.8237

76.2785 19.0160
5.8243
74.7602 20.5172
5.8224

Chapter 6

Table 6.9. Results for manufacturer and retailers under different component cost
Manufacturer &
Retailer

Arl
75

Results
Price
Cycle
5

Demand(10 )
7

100

175

125

Profit(10 )
Price
Cycle
Demand(105)
Profit(107)
Price
Cycle
5

Demand(10 )
7

Profit(10 )

NCS case

WCS case

m
r1

r2

789.2185

1,061.7

0.0422

0.0422

5.1359

5.2647

Product 1

Product 2

694.8654

5.3383
694.2187
793.4469
0.0430
0.0430
5.1405
5.2393
5.1376

r1

r2

766.7627

1,061.9

1,216.7

0.0425

0.0425

0.0425/11

0.0425/18

5.1339

5.3994

5.1339

5.3994

18.826
1,061.9
0.0425/11
5.1339
18.826

24.295
1,216.7
0.0425/18
5.3994
24.295

Product 1

Product 2

1,227.9

695.1555

0.0422/11

0.0422/18

5.1359

5.2647

18.841
1,061.4
0.0430/11
5.1405
18.874

23.097
1,230.1
0.0430/18
5.2393
22.875

5.0500
695.1556
766.7619
0.0425
0.0425
5.1339
5.3994
5.0501

693.4098

798.7334

1,061.0

1,232.7

695.1703

766.7820

1,061.9

1,216.7

0.0442

0.0442

0.0442/11

0.0442/18

0.0425

0.0425

0.0425/11

0.0425/18

5.1461

5.2076

5.1461

5.2076

5.1338

5.3993

5.1338

5.3993

18.916

22.599

18.825

24.294

4.8898

5.0512

Chapter 6

CHAPTER 7
INTEGRATED SUPPLIER SELECTION, PRICING AND INVENTORY
DECISIONS IN A MULTI-LEVEL SUPPLY CHAIN

7.1 INTRODUCTION

Many firms identify and qualify adequate suppliers to provide the materials and
service needed by them (Watts and Hahn 1993). In the previous chapter, we integrated
supplier selection into pricing and inventory coordination problem. However, the
effective evaluation of these qualified suppliers is also an important task faced by
many industrialists (Morgan 1996; Poirier 1999; Michael and Chong 2001). This
chapter puts forward a configuration model for optimal supplier selection, pricing and
inventory decisions and proposes a cooperative game theory approach to evaluate the
suppliers for integrated multi-level supply chain.
The literature shows that to maximize overall system profit and integrate
operations among various entities through coordination are of best interest. For
example, Alchian and Demsetz (1972) and Jensen and Meckling (1976) point that
vertical integration would exploit synergies between different divisions appeared as
economies of scope. Williamson (1975) looks at the long-term relationship between a
seller and buyer and finds the advantage of integration, for example, saving
transaction cost.
The integration of supply chain mainly focuses on the pricing model, distribution

176

Chapter 7

inventory model. Boyaci and Gallego (2002) analyze the problem of integrating
pricing and inventory replenishment policies in a supply chain consisting of a
wholesaler, one or more geographically dispersed retailers. They show that optimally
coordinated policy could be implemented cooperatively by an inventory-consignment
agreement. Moutaz (2003) develops inventory coordination mechanism for the supply
chain composed of a single supplier, three manufacturers and two retailers per
manufacturer. Jabber and Goyal (2008) consider optimal order quantity model in a
multiple suppliers, a single vendor and multiple buyers supply chain. Some
researchers also consider supplier selection decisions in the vertical integration model.
Haq and Kannan (2006) propose an integrated model to select best suppliers and
optimize inventory level for a multi-level supply chain to minimize the total supply
chain cost. Liao and Rittscher (2007) develop a multi-objective programming model
integrating supplier selection, procurement lot sizing and carrier selection decisions
under dynamic demand condition. However, none of them incorporates the above
three aspects, i.e., pricing, inventory, and supplier selection decisions.
Papers employing cooperative game theory to study supply chain management are
becoming more popular (Nagarajan and Soi 2006). Gerchak and Gupta (1991)
model their inventory problem as a cooperative game and analyze the joint cost
allocation problem. Granot and Soi (2003) study a multi-stage model of
decentralized distribution system with inventory sharing consisting of multiple
retailers. In the cooperative stage, they show that Shapley value encourages the

177

Chapter 7

retailers to share all of their residuals. Edward (2008) constructs a cooperative supply
chain game for calculating transfer prices for intermediate goods in vertically
integrated supply chain. Leng and Parlar (2009) model a three-level supply chain
cooperative game in characteristic-function form to allocate cost saving among supply
chain members.
This chapter considers three-level supply chain composing of multiple suppliers,
one manufacturer and multiple retailers as described in the last chapter. The
manufacturer obtains components from the potential qualified suppliers and produces
final products for retailers in different markets. All the suppliers, the manufacturer and
the retailers would like to be involved in an integrated manner to work on the pricing,
inventory and supplier selection decisions to maximize the overall profit of the supply
chain system. There are several research questions to be investigated:
(Q7.1) what are the optimal retail prices for products sold to different markets?
(Q7.2) how frequent should the production process take place?
(Q7.3) what are the optimal suppliers and components that should be selected?
(Q7.4) what is importance of each supply chain member to the supply chain in an
integrated process?
We formulate a mixed-integer programming model to integrate the optimal
supplier selection, pricing and inventory decisions. A cooperative game theoretic
approach is proposed and the Shapley value is used to evaluate the supply chain
members for the above supply chain in an integrated process. Their use is tested

178

Chapter 7

through a numerical example. The impacts of the market scale parameter, setup cost
and component cost on the optimal decisions and the marginal contributions are
studied.
The reminder of this paper is organized as follows. Section 7.2 describes the
problem. In Section 7.3, we formulate the mathematical models and solution
procedure. A numerical study and the influence of market scale parameter, component
cost and setup cost are presented in Section 7.4. Finally, this chapter concludes in
Section 7.5.

7.2 PROBLEM DESCRIPTION

The supply chain context follows that of Chapter 6. We concern a three-echelon

supply chain involving multiple suppliers, one single manufacturer and multiple
retailers. The manufacturer, indicated by m , designs and customizes a set of products
for retailers ( rl , l 1, 2,..., L ) in different independent market segments. Each retailer
is served by one type of product.
The architecture for all the product is finalized which consists of a series of
different functionality elements. The components offering different levels of the same
functionality are grouped together in the substitutable component set (SCS), indexed
by i 1, 2,..., I . The components within the same SCS can be ranked in order of
decreasing functionality and higher functionality components can substitute ones with
lower functionality completely, but not vice versa. Let N i be the number of

179

Chapter 7

components in SCS i . Lij is used to denoted the component which is the j th

element in the i th SCS, where i 1, 2,..., I and j 1, 2,..., Ni . All the components
are obtained from a fixed number of alternative qualified suppliers ( sv , v 1, 2,...,V ).
All the suppliers, the manufacturer and the retailers are supposed to be willing to
involve in an integrated process of the supply chain, that is the chain members would
like to vertical integrate to maximize the overall supply chain system profit. Thus, all
the supply chain members jointly determine the optimal production setup time
interval and retail prices, and the best suppliers and components to maximize the
overall supply chain system profit.
Prior to develop the mathematic model for solving the above problems, we make
the following underlying assumptions and simplifications:
(A7.1): The demand at retailer l is almost invariably a downward sloping and convex
function with respect to the retail price in real world. We employ the linear demand
function.
(A7.2): Vendor managed inventory (VMI) system is employed between the suppliers
and the manufacturer and between the manufacturer and the retailers to replenish the
components and the products. Under this assumption, the inventories of the final
products are on the manufacturers side and the retailers take zero inventory. The
components inventories for the manufacturer and the suppliers are at the side of their
upstream suppliers and their corresponding inventory holding costs are also borne by
their upstream suppliers.

180

Chapter 7

(A7.3): Single sourcing strategy is adopted between supplies and manufacturer. Thus,
the manufacturer obtains one type of component from only one supplier.
(A7.4): Either all or none the demand of a component is replaced by the lower order
component in the same SCS.
(A7.5): Shortage are not permitted, hence the annual production capacity is greater
than or equal to the total annual market demand.

7.3 MATHEMATICAL MODELS FORMULATION

7.3.1 Notations
In this section, we will introduce some other relevant parameters and decisions
variables used for the mathematical models.
Parameters:

Drl : Retailer ls annual demand

Arl : A constant in the demand function of retailer l, which represents his market
scale

erl : Coefficient of the products demand elasticity for retailer l

Pl : Manufacturers annual production capacity for product l

cml : Production cost per unit product l

hmpl : Holding costs per unit of product l inventory
S m : Setup cost per production

FLij : Fixed cost of using component Lij

181

Chapter 7

uLik l : Usage amount of unit component Lij per unit product l

wsv : The fixed cost of using supplier v, covering supplier certification, contract
setup, etc.

sv : Number of different types of components supplier v is capable of supplying

cLij : The cost of component Lij paid by supplier

Rrl : Retailer ls annual fixed costs for the facilities and organization to carry this
product

Rm : Annual fixed costs for the facilities and organization for the production of the
products

Rsv : Supplier vs annual fixed costs for the facilities and organization to carry the
components
Decision variables:

prl : Retail price charged to the customer by retailer l

T : Decision variable, manufacturers setup time interval

ijk : Binary decision variable to indicate whether component Lij has been used to
replace Lik

s : Binary decision variable to indicate whether supplier v is used

v

r : Binary decision variable to indicate whether retailer l is used

l

tsv Lij : Binary decision variable to indicate whether component Lij is supplied by
supplier v

z Lij : Binary decision variable to indicate whether component Lij is used

182

Chapter 7

7.3.2 Integrated supply chain model

The objective function of the integrated supply chain model is concerned with the
maximization of the overall profit of the supply chain. The total cost of the supply
chain is composed of the production costs, fixed costs to contract with the suppliers
and using components, holding costs for the products, setup and ordering costs,
component costs and annual fixed costs for all the chain members. According to the
given VMI policy, only the manufacturer has to pay for inventory for final products.
The behavior of his inventory level for the product l is illustrated as Figure 7.1. The

hmpl Pl Drl TDrl

annual inventory for product ls is given by

(suggested by Zhang

2 Pl

(2007)). The setup cost S m and ordering cost Om occurs at the beginning of each
production. Thus, the integrated supply chain model (A) is formulated as the
following constrained mixed-integer programming problem:
L

Ni

l 1

l 1

v 1

i 1 j 1

ij

ij

hmpl Pl Drl TDrl

l 1

2 Pl

max Z pr Dr r cm Dr r ws s FL zL

Sm Om
T

Ni
V
L
I Ni
L
V

cLij z Lij uLij l uLik l ijk Drl tsv Lij Rrl Rm Rsv sv
v 1 l 1 i 1 j 1
k j 1
l 1
v 1

(7.1)
Subject to:

Drl Arl erl prl

(7.2)

0 Drl Pl

(7.3)

k 1

j 1

ijk

z Lik 1 , uLik l 0 , i 1, 2,..., I ; j 1, 2,..., Ni 1; k j 1,..., Ni ; (7.4)

ijk zL , i 1, 2,..., I ; j 1, 2,..., Ni 1; k j 1,..., Ni

(7.5)

tsv Lij 0 , Lij Qsv , v 1, 2,...,V

(7.6)

ij

183

Chapter 7

sv Lij

tsv Lij sv sv , v 1, 2,...,V

v 1

Lij Qsv

zLij , i 1, 2,..., I ; j 1, 2,..., Ni

(7.8)

zLij , tsv Lij , sv , ijk , rl 0,1 , i 1, 2,..., I ; j 1, 2,..., Ni 1; k j 1,..., Ni

prl 0

T 0.

(7.7)

l 1, 2,..., L

(7.9)
(7.10)

Constraint (7.3) gives the bounds of the annual demand, which cannot exceed the
annual production capacity Pl of the product. As indicated in the fourth point of the
assumption in Section 7.2, constraint (7.4) ensures that for the component predefined
for the products, it is either used or replaced by higher functionality one, but not both.
Constraint (7.5) makes sure that only procured components can be used to replace
other components. (7.4) and (7.5) together ensure that the demands for all components
are satisfied. Also, they met the one-way substitutability constraint which ensures that
a higher functionality component can replace a lower functionality component but not
vice versa. Constraint (7.6) sets the value of tsv Lij 0 for all components Lij Qsv for
all the suppliers. Constraint (7.7) indicates that a component is procured from exactly
one supplier. Constraint (7.8) sets the value of sv to 1, if supplier v supplies a
component, and ensures that the number of different types of components supplied by
supplier v is no greater than sv . The value ranges of all the variables are set by
constraints (7.9) and (7.10).

184

Inventory level

Chapter 7

Pl Drl

Drl

Time

Figure 7.1. Manufacturers inventory level for final product l

7.3.3 Cooperative game model formulation

In this section, we use a cooperative game to evaluate the suppliers involved in the
above supply chain. To formulate a cooperative game, we must first identify the set of
all players, N, and a value function v , that associate with every nonempty subset G
(a coalition) of all the players (Osborne and Rubinstein, 2002). A real number v G
represents the worth of G that describes the total expected gain from the coalition
G . And v satisfies the following two conditions: (1) v 0 , and (2) If

R, T N and R T , then v R T v R v T .
Based on the definition, we define that the set of the game players is composed of
the suppliers, the manufacturer and the retailers, i.e., V L 1 supply chain
members. The supply chain is an integrated process, wherein the components are
manufactured to final products, then delivered to customers (Benita 1999). Thus, to
formulate cooperative game for such supply chain in an integrated process, the value
function, here, we regard it to be,

185

Chapter 7

Z G , if coalition G could transfer components to final products

, (7.11)
0,
otherwise

where Z G is the integrated model (A) with the constraints that the selection
variables for those supply chain members out of coalition G to be zero. For example,
if supplier 1 is not within coalition G , Z G is the model (A) adding constraint

s 0 . We can see that a coalition G is valid if and only if G could transfer

1

components to one product at least. Obviously, the value function defined above
satisfies conditions (1). For condition (2), R, T V L 1 and R T , if the
manufacturer

m R, T , we have R T R T 0 , because the

manufacturer is the indispensable supply chain member to transfer components to

final products; otherwise, without loss of generality, we assume m R , thus,

T 0 , Hence. R T R T . Thus, ( V L 1 , ) is a cooperative

supply chain game.
The Shapley value is a solution concept for cooperative game, which is defined as
follows:

i N , v

S ! n S 1!

S N \i

n!

v S

i v S ,

(7.12)

where i N , v is the Shapley value of player i , S is the players number in the

coalition S , n is all players numbers, and v S is the value function of the
coalition S .
The

v S

marginal

i v S .

contribution

of

adding

player

to

coalition

is

According to the definition, the Shapley value is obtained by

186

Chapter 7

averaging the marginal contributions for all the coalitions of the game players
(Shapley, 1971). In this chapter, we use the Shapley value to evaluate the supply chain
members, especially the suppliers. Their importance to the integrated supply chain is
classified according to their marginal contributions.
7.3.4 Solution procedure
The integrated supply chain model (A) has binary variables and L 1
non-integer variables. Some researchers solve such problems through genetic
algorithms and conduct the comparison with the results from Lingo software (Peters
and Maheswaran 1996; Tavakkoli and Rahimi 2006). In this chapter, we simply
employ Lingo for the solving the integrated models. To find the Shapley values for
each supply chain member, we use the following procedure:
Step 1. Denote R r1 , r2 ,..., rL to be the set of all the retailers; identify all the
subset of R . For each X , X R , find out all the suppliers combination S X
satisfying constraints (7.4)-(7.9) in Model (A) and all the valid coalitions

GX S X i , m, X , S X i S X .

Step 2. Calculate the value function Z GX t , GX t GX , X R , using Lingo.

Step 3. Use (7.12) to calculate the Shapley values for all the supply chain members.

7.4 NUMERICAL RESULTS

To demonstrate the implication of the proposed method and solution procedure,

we present and discuss a specific case in a computer industry. This particular case is

187

Chapter 7

concerned with optimal supplier selection, pricing and production decisions and
evaluation of the suppliers in the supply chain (as Figure 6.1) described in Chapter 6.
Through this example, we also conduct sensitivity analysis on market scale parameter,
component cost and production setup cost. The initial numerical parameters are
presented as the previous chapter.
7.4.1 Results for base example
The results for Model (A) for the base example are calculated by Lingo as Table
7.2 shows. We then calculate the Shapley values for all the chain members using the
procedure in Section 7.3. In this research, the Shapley value is simply employed to
study the marginal contribution of the supply chain members instead of allocating the
profits. For the convenience of comparison, we consider the value of grand coalition
to be 1, thus, the values for the other coalitions are the ratio of the profit of this
coalition to the profit of the grand coalition. Using the solution procedure in Section
7.3.4, the Shapley values can be obtained as Table 7.2. For example, for supplier 1,
his Shapley value is:

s
1

SC

S 1!V L 1 S !
V L 1!

Z S Z S \ s / Z ,
1

(7.13)

where C is the set of all the valid coalitions shown in Table 7.1; Z S is the profit
of coalition S in Table 7.1; Z is the profit of this grand coalition.
From Table 7.2, we can see that supplier 4 is not selected for any component and
supplier 6 is selected for component L31 . Component L21 supplied by supplier 1 is
not used for any product. The retail price and demand for product 2 are higher than
188

Chapter 7

product 1 for retailer 1.

Supplier 5 has the largest Shapley value. This is because supplier 5 is the one and
only supplier for the indispensable component L51 for the final products. Although
supplier 1 only provides components for retailer 1, his Shapley value is the second
largest one of the suppliers, since the component L11 he supplied cannot be
substituted. Supplier 4 gets the lowest Shapley value. Retailer 2 has larger Shapley
value than retailer 1 for his better market environment.
7.4.2 Sensitivity analysis
In this section, we conduct sensitive analyses on market scale parameter,
component cost and production setup cost to study the influence of these parameters
on the optimal decisions and marginal contributions of the supply chain members.
7.4.2.1 Influence of market scale parameter
To examine the influence of the market scale, we experimented on the proposed
model by changing the retailer 1s market scale parameter. Table 7.2 summarizes the
results of sensitivity analysis with two different market scales of retailer 1.
The experiments indicate that as the market scale of retailer 1 increases, the price
and demand for product 1 increase. For example, when Ar1 increases from 2.0106
in the base example to 3.5106, the retail price for product 1 increases from
924.2871 to 1460.001, by 57.96% and the demand increases to 14.55999105, by
106.23%. However, retailer 2 would not change his retail price for product 2. For the
manufacturer, he will short up his setup time interval. That is the manufacturer sets his

189

Chapter 7

production more frequently. For example, in the above case, the setup cycle decreases
from 0.0152 to 0.0136, by 10.53%.
Next, we present the changes of the Shapley values for all the suppliers and
retailers under each level of Ar1 as Figure7.2. We can see that the Shapley values for
supplier 1 and retailer 1 increase most significantly. For supplier 2, supplier 3 and
retailer 2, their Shapley values decrease. This makes the Shapley values for supplier 4
and supplier 6 relatively larger to the entire supply chain, although these values have
no significant change. So we can see that when one retail market improves, the
importance for the retailer increases and those suppliers who provide components not
used for the product in this market will become less important.
7.4.2.2 Influence of production setup cost
As Table 7.2 shows, the production setup cost has very slight influence on the
pricing decisions. As S m increases from 100 to 250, the retail price for product 1
lightly increases from 924.2871 to 924.2879, and for product 2 as well. However, its
influence on the setup time interval for the manufacturer is significant. For example,
when S m increases from 100 to 250, the setup time interval is lengthen to 0.024, by
57.89%. The change of S m also has very limited influence on the Shapley values of
the supply chain members.
7.4.2.3 Influence of component cost
A higher component cost will result in a higher cost for product using this
component. Considering such effect, it would be worthwhile examining the interplay

190

Chapter 7

between component cost and optimal supply chain decisions. As such, we experiment
on the model with varying costs of component L11 .
When the cost of component L11 raises from 100 to 200, the retail price for
product 1 increases from 924.2871 to 974.2871, by 5.41% and its demand decreases
from 7.05998105 to 6.35998105, by 9.92%. The product 2s price and demand are
almost the same with the base example. The manufacturer sets his production less
frequent. In the above case, the setup time interval increases from 0.0152 to 0.0154,
by 1.32%.
(a) Shapley value for suppliers

0.25

Shapley value

Shapley value

0.25
base example
Ar1:3500000
Ar1:5000000

0.2

(b) Shapley value for manufacturer and retailers

base example
Ar1:3500000
Ar1:5000000

0.2

0.15

0.15
0.1

0.1

0.05

0.05
0

0
s1

s2

s3

s4

s5

s6

Supplier

r1
r2
Manufacturer & retailers

Figure 7.2. Shapley value for different market scale parameters

(a) Shapley value for suppliers

(b) Shapley value for manufacturer and retailers

0.25

Shapley value

Shapley value

0.25
base example
cL11:200
cL11:300

0.2

base example
cL11:200
cL11:300

0.2

0.15

0.15
0.1

0.1

0.05

0.05
0

0
s1

s2

s3

s4

s5

s6
Supplier

r1
r2
Manufacturer & retailers

Figure 7.3. Shapley value for different component costs

Figure 7.3 presents the changes of the Shapley values for the suppliers and the
retailers under different costs of component L11 . The Shapley values for supplier 1
and retailer 1 decrease as the increase component L11 s cost . For supplier 2, supplier
191

Chapter 7

3 and retailer 2, the results are reversed. That is to say when one products component
cost increases, the component supplier and the corresponding retail market for this
product will become less important to the supply chain system. This is opposite to the
results from the increase of market scale.
7.4.3 Managerial implications
Based on the results obtained from the simulation results presented above, we can
see that the proposed cooperative game can be used to evaluate the supply chain
members by calculating their Shapley values. We wrap up our findings and some
implications regarding the impacts of the above parameters analyzed above.
Firstly, the increase of one retailers market scale will increase the product price
and demand in this market without having impacts on the other retail market. This is
because the entire supply chain could benefit from the higher price and larger demand
from the increase of one retailers market while maintaining the profit level in the
other retail market.
Secondly, the retailer gains an increased importance in the supply chain as his
market scale increases and the importance for the suppliers who provide components
not used for the product in this market will reduce. This is due to the marginal
contributions of the retailer and his component suppliers as the retailers market scale
increases.
Thirdly, the increase of one products component cost would not influence the
product in other retail market which does not use this component. Meanwhile, this

192

Chapter 7

product retailer and the component supplier might become less important to the entire
supply chain. A larger component cost will result in lower profit from the product that
uses this component, and thus the lower marginal contributions of the retailer and the
supplier obtains. Besides, the increase of component cost will not increase the product
cost for other retail market and the entire supply chain could not obtain larger profit
from the other market since the profit from the other market is maximal.
Lastly, the change of setup cost will significantly influence the setup time interval
for the manufacturer. However, this cost change has relatively little impact on the
pricing decisions and marginal contributions of the supply chain members. The reason
is that the manufacturer could counteract the variation of setup cost mainly through
changing his setup time interval, and thus the effect of setup cost change on the profit
of the supply chain is very small.

7.5 SUMMARY

This chapter proposes an integrated supply chain model for optimal supplier
selection, pricing and inventory decisions in a multi-level supply chain. A cooperative
game approach is used to evaluate the supply chain members. A numerical study is
conducted to examine the integrated supply chain model and cooperative game model.
We also analyze the impacts of market scale parameter, component cost and setup
cost.
The results of the numerical analysis show that: (a) the increase of one retailers

193

Chapter 7

market scale will advance the retail price and demand of the product in this market,
but this increase will have no effect on the other retail market; an increased market
scale enhances the importance of this market and reduces the importance of the
suppliers provided components not used for the product in this market.; (b) the
increase of one products component cost will not affect the other product retail
market; this product retailer and the component supplier will become less important to
the entire supply chain which is just the opposite results to the increase of the
retailers market scale; (c) the change of setup cost has relatively little impact on the
pricing decisions and marginal contributions of the supply chain members.

194

Table 7.1. Profits for all valid coalitions

Coalitions

Base example

Ar1 3500000

Ar1 5000000

cL11 200

cL 300
11

Sm 250

Sm 400

195

s6, m, r1}
s6, m, r1}
s6, m, r1}
m, r1}
s6, m, r1}
s6, m, r1}
m, r1}
s6, m, r1}

0.2013933E+09 0.1172105E+10 0.2946389E+10 0.1517937E+09 0.1091942E+09 0.1091942E+09 0.1091942E+09

0.3351525E+09 0.1470864E+10 0.3410149E+10 0.2701529E+09 0.2121533E+09 0.2121533E+09 0.2121533E+09
0.3351525E+09 0.1470864E+10 0.3410149E+10 0.2701529E+09 0.2121533E+09 0.2121533E+09 0.2121533E+09
0.3420375E+09 0.1485249E+10 0.3432034E+10 0.2763379E+09 0.2176383E+09 0.2176383E+09 0.2176383E+09
0.3560174E+09 0.1514229E+10 0.3476014E+10 0.2889177E+09 0.2288181E+09 0.2288181E+09 0.2288181E+09
0.3560174E+09 0.1514229E+10 0.3476014E+10 0.2889177E+09 0.2288181E+09 0.2288181E+09 0.2288181E+09
0.3420375E+09 0.1485249E+10 0.3432034E+10 0.2763378E+09 0.2176383E+09 0.2176383E+09 0.2176383E+09
0.3560174E+09 0.1514229E+10 0.3476014E+10 0.2889177E+09 0.2288181E+09 0.2288181E+09 0.2288181E+09

s2, s3,
s5,
s2, s3, s4, s5,
s2, s3, s4, s5,
s4, s5,
s3,
s5,
s3, s4, s5,
s2,
s4, s5,
s2,
s4, s5,
s2, s3,
s5,
s2, s3, s4, s5,
s2, s3, s4, s5,

s6, m,
m,
s6, m,
s6, m,
s6, m,
s6, m,
m,
s6, m,
s6, m,
m,
s6, m,

r2}
r2}
r2}
r2}
r2}
r2}
r2}
r2}
r2}
r2}
r2}

0.5200721E+09 0.5200721E+09 0.5200721E+09 0.5200721E+09 0.5200721E+09 0.5200721E+09 0.5200721E+09

0.5043921E+09 0.5043921E+09 0.5043921E+09 0.5043921E+09 0.5043921E+09 0.5043921E+09 0.5043921E+09
0.5200721E+09 0.5200721E+09 0.5200721E+09 0.5200721E+09 0.5200721E+09 0.5200721E+09 0.5200721E+09
0.2979930E+09 0.2979930E+09 0.2979930E+09 0.2979930E+09 0.2979930E+09 0.2979930E+09 0.2979930E+09
0.4440723E+09 0.4440723E+09 0.4440723E+09 0.4440723E+09 0.4440723E+09 0.4440723E+09 0.4440723E+09
0.4440723E+09 0.4440723E+09 0.4440723E+09 0.4440723E+09 0.4440723E+09 0.4440723E+09 0.4440723E+09
0.3477528E+09 0.3477528E+09 0.3477528E+09 0.3477528E+09 0.3477528E+09 0.3477528E+09 0.3477528E+09
0.3607927E+09 0.3607927E+09 0.3607927E+09 0.3607927E+09 0.3607927E+09 0.3607927E+09 0.3607927E+09
0.5200721E+09 0.5200721E+09 0.5200721E+09 0.5200721E+09 0.5200721E+09 0.5200721E+09 0.5200721E+09
0.5043921E+09 0.5043921E+09 0.5043921E+09 0.5043921E+09 0.5043921E+09 0.5043921E+09 0.5043921E+09
0.5200721E+09 0.5200721E+09 0.5200721E+09 0.5200721E+09 0.5200721E+09 0.5200721E+09 0.5200721E+09

s2, s3,
s5,
s2, s3, s4, s5,
s2, s3, s4, s5,
s4, s5,
s3,
s5,
s3, s4, s5,
s2,
s4, s5,
s2,
s4, s5,
s2, s3,
s5,
s2, s3, s4, s5,
s2, s3, s4, s5,

s6, m,
m,
s6, m,
s6, m,
s6, m,
s6, m,
m,
s6, m,
s6, m,
m,
s6, m,

r1,
r1,
r1,
r1,
r1,
r1,
r1,
r1,
r1,
r1,
r1,

1.
2.
3.
4.
5.
6.
7.
8.

{s1,
{s1,
{s1,
{s1,
{s1,
{s1,
{s1,
{s1,

s2,
s2,
s2, s3,
s2, s3,
s2, s3,

9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.

{
{
{
{s1,
{s1,
{s1,
{s1,
{s1,
{s1,
{s1,
{s1,

20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.

{
{
{
{s1,
{s1,
{s1,
{s1,
{s1,
{s1,
{s1,
{s1,

s3,
s3,

s4, s5,
s5,
s4, s5,
s4, s5,
s4, s5,
s5,
s4, s5,
s4, s5,

r2}
r2}
r2}
r2}
r2}
r2}
r2}
r2}
r2}
r2}
r2}

0.5200721E+09 0.5200721E+09 0.5200721E+09 0.5200721E+09 0.5200721E+09 0.5200721E+09 0.5200721E+09

0.5043921E+09 0.5043921E+09 0.5043921E+09 0.5043921E+09 0.5043921E+09 0.5043921E+09 0.5043921E+09
0.5200721E+09 0.5200721E+09 0.5200721E+09 0.5200721E+09 0.5200721E+09 0.5200721E+09 0.5200721E+09
0.4993916E+09 0.1470104E+10 0.3244389E+10 0.3929921E+09 0.2995928E+09 0.4993846E+09 0.4993796E+09
0.7792305E+09 0.1914943E+10 0.3854228E+10 0.6442310E+09 0.5222315E+09 0.7792230E+09 0.7792175E+09
0.7792305E+09 0.1914943E+10 0.3854228E+10 0.6442310E+09 0.5222315E+09 0.7792230E+09 0.7792175E+09
0.6582460E+09 0.1801459E+10 0.3748244E+10 0.5925462E+09 0.5338465E+09 0.6582387E+09 0.6582335E+09
0.6846659E+09 0.1842879E+10 0.3804664E+10 0.6175661E+09 0.5574663E+09 0.6846585E+09 0.6846532E+09
0.8373453E+09 0.1995558E+10 0.3957343E+10 0.7702455E+09 0.7101457E+09 0.8373377E+09 0.8373321E+09
0.8082854E+09 0.1951498E+10 0.3898283E+10 0.7425700E+09 0.6838858E+09 0.8082778E+09 0.8082723E+09
0.8373453E+09 0.1995558E+10 0.3957343E+10 0.7702455E+09 0.7101457E+09 0.8373377E+09 0.8373321E+09

Chapter 7

Table 7.2. Decisions and Shapley values

(a) Decisions and Shapley values for suppliers
s1
Suppliers
Parameter

Results
Component
Base
example Shapley

3.510

Ar1

5.010

196

bL11

Sm

L22

L41

L31

L41

L51

L31

L61

0.106830054

0.05538969

0.074053153

0.044060008

0.224682872

0.052693113

0.16818855

0.040170719

0.046301836

0.045387806

0.217643386

0.049994538

0.19053441

0.033628624

0.035862446

0.046121855

0.215472856

0.048941929

0.095798

0.069773858

0.077490878

0.043636737

0.223928605

0.051162908

0.084142

0.085106479

0.081049565

0.043209443

0.223109145

0.049502518

0.106831449

0.055389839

0.074053754

0.044061076

0.224688319

0.052693197

0.106830969

0.055390209

0.074054249

0.044061371

0.224689821

0.052693549

Component
Shapley

400

L61

s6

Component
Shapley

250

L23

s5

Component
Shapley

300

L12

s4

Component
Shapley

200

L21

s3

Component
Shapley

L11

s2

Component
Shapley

: implies the component is not selected for the final products

Chapter 7

Chapter 7

(b) Decisions and Shapley values for manufacturer and retailers

Manufacturer
& retailers
Parameter

0.224682872

Shapley

3.510

1033.335

7.05998

7.59998

0.081918367

0.135664793

1460.001

1033.335

14.55999

7.59998

0.15773533

0.056926434

1995.714

1033.335

22.06000

7.59998

0.185263193

0.028706129

974.2871

1033.341

6.35998

7.59990

0.068715754

0.145576962

1024.287

1033.335

5.65998

7.59998

0.054767634

0.155996157

924.2879

1033.336

7.05997

7.59996

0.081920353

0.135668082

924.2884

1033.337

7.05996

7.59995

0.081919707

0.13566899

0.0136

Setup cycle
Price
5

Demand(10 )

0.217643386

Shapley

5.010

0.0130

Setup cycle
Price
5

Demand(10 )

0.215472856

Shapley

200

0.0154

Setup cycle
Price
5

Demand(10 )

0.215472856
0.0156

Shapley

300

Setup cycle
Price
5

Demand(10 )

0.223109145
0.0240

Shapley

250

Setup cycle
Price
5

Demand(10 )

Sm

924.2871

0.0152

Price
Base
example Demand(105)

bL11

r2

Results
Setup cycle

Ar1

r1

0.224688319

Shapley

400

0.0303

Setup cycle
Price
5

Demand(10 )
Shapley

0.224689821

197

CHAPTER 8
CONCLUSION AND FUTURE RESEARCH

8.1 MAJOR ACHIEVEMENTS AND CONTRIBUTIONS

This thesis aims to coordinate the supply chain configuration decisions in

multi-level supply chains by game-theoretic approaches. The investigation of this
research is mainly composed of four sub-problems or scenarios: (I) CCSC with
pricing decision, (II) CCSC with competing retailers, (III) CCSC with price and
inventory coordination, (IV) CCSC with supplier selection. These four parts constitute
the major achievements of this research. In addition, we also attempt to study the
supply chain configuration problem from an integrated and cooperative viewpoint.
Table 8.1 summarizes the major achievements of in this thesis. In the following
sections, we conclude the achievements made for each chapter in detail.
8.1.1 CCSC with pricing decision

Chapter 3 discusses price coordination problem in a three-level supply chain

composed of a single supplier, a single manufacturer and a single retailer. The
supplier provides the manufacturer with the sole raw material used to produce a single
product sold to the retailer. The chain members make their corresponding optimal
pricing decisions under different channel and power structures. We consider three
different channel structures for this supply chain. The first is the decentralized channel

198

Chapter 8

that the manufacturer uses independent supplier and retailer, in which they optimize
their own profit individually and non-cooperatively. The second semi-integrated
channel is that the manufacturer backward integrates with the supplier or forward
integrates with the retailer and uses independent retailer or supplier simultaneously.
The third one is the integrated channel structure, where the manufacturer, the supplier
and the retailer operate in total cooperation in a full vertical integration system. Two
power structures (i.e., leader-follower, independent) are studied for the decentralized
and the semi-integrated channels. The leader-follower power structure is modeled as a
Stackelberg game, where the manufacturer always takes the leadership and makes the
first move in the game. The independent power structure is treated as a Nash game, in
which the supply chain members make their pricing decisions simultaneously and
non-cooperatively.
We further investigate the effects of power structures, channel structures and
market related parameters on the optimal prices and profits for the individual supply
chain members and the entire system as well. The results show that the manufacturer
or the integrated members had better take the channel leadership for either
decentralized or semi-integrated channel, the equilibrium prices are lower and profits
are higher compared with the channel without such leadership. The integration for the
manufacturer and the retailer cannot always improve their profits in a monopoly
unless the price elasticity b is no less than a certain level (3.5396). When product cost
is larger than a certain level (0.2749), the chain members profits will increase as the

199

Chapter 8

market becomes more sensitive to the retail price.

Table 8.1. Summary of CCSC problem in this thesis
Scenario

Configuration
Decisions

Supply Chain
Network

(I).CCSC
with pricing
decision

Pricing

Single member in
each stage

(II).CCSC
with
competing
retailers

Pricing,
inventory

One manufacturer,
multiple competing
retailers

(III).CCSC
with price
and
inventory
coordination
(IV).CCSC
with
supplier
selection
(V).CCSC
with
integration

Pricing,
inventory

Multiple suppliers,
one manufacturer,
multiple retailers

Pricing,
inventory,
supplier
selection
Pricing,
inventory,
supplier
selection

Multiple alternative
suppliers,
one manufacturer,
multiple retailers
Multiple alternative
suppliers,
one manufacturer,
multiple retailers

Major Achievements
Investigate and compare optimal
pricing strategies under different
channel structures and power
structures
Formulate game model for
two-level supply chain
Determine optimal pricing and
inventory decisions for supply
chain members
Formulate game model for
multi-level supply chain
Determine optimal pricing and
inventory decisions for supply
chain members
Simultaneous incorporate supplier
selection into optimal pricing and
inventory decisions model for
multi-level supply chain
Formulate
an
mixed-integer
programming to determine supply
chain configuration decisions
Use cooperative game to evaluate
the members of a supply chain in
an integrated process

8.1.2 CCSC with competing retailers

In this chapter, we consider how a manufacturer interacts with his multiple

competing retailers to optimize their individual net profits through coordinating
pricing and inventory decisions in a two-echelon supply chain. The manufacturer
purchases raw materials outside to produce substitutable products to different retailers.
200

Chapter 8

The manufacturer determines his optimal wholesale prices for different retailers, setup
time interval for the products and replenishment cycle for raw materials to maximize
his profit. For each retailer, he considers optimal retail prices and replenishment
policies to maximize his profit. The problem is modeled as a dynamic
non-cooperative game in which all the competing retailers formulate a subgame and
as a sector play the whole game with the manufacturer. Analytical method and
solution procedure are developed to determine the Nash equilibrium of this game. A
numerical study is conducted to examine the influence of different parameters on
decisions and profits of the supply chain and its constituent members. Several useful
managerial implications are generated and discussed.
The numerical results show that: firstly, when one retailers market scale becomes
larger or more sensitive to his price or his product becomes more substitutable for
other product, his profit will be increased, while the other retailers profits will
decrease; Secondly, the increase of retailers holding cost will lengthen the
manufacturers setup time interval, while shorten the retailers replenishment cycle
time; thirdly, the change of manufacturers production capacity or setup cost has
mainly impact on the manufacturers setup time interval; lastly, the increase of the
manufacturers holding cost may increase the profits of the retailers.
8.1.3 CCSC with price and inventory coordination

Extending the work presented in Chapter 4, Chapter 5 examines coordination of

pricing and inventory decisions in a multi-level supply chain composed of multiple
201

Chapter 8

suppliers, a single manufacturer and multiple retailers. The retailers face the customer
demands of different products, which can be produced by the manufacturer with
different raw materials purchased from the suppliers. The suppliers, the manufacturer
and the retailers determine their optimal pricing and inventory decisions to maximize
their individual net profits. The problem is modeled as a three-level nested Nash game
in which all the suppliers formulate the bottom-level Nash game, the whole supplier
sector play the middle-level Nash game with the manufacturer, and both sectors as a
group player formulate the top-level Nash game with the retailers. Analytical method
and solution algorithm are developed to determine the equilibrium of the game.
The model and proposed solution procedure are illustrated through a simulation
experiment and a series of sensitivity analyses conducted to understand the influence
of market, production and raw material related parameters on decisions and profits of
the supply chain and its constituent members. Experiment results show that: when one
retailers market becomes more sensitive to the price, his profit will decrease, but the
other retailers profit will increase; suppliers profits may increase as the
manufacturers setup cost increases; when one suppliers raw material cost increases,
the other suppliers profits might decrease more significant than his own; the setup
time interval for the manufacture will be lengthened if the retailers price sensitivity,
the manufacturers setup cost or the suppliers raw material cost increases.
8.1.4 CCSC with supplier selection

Chapter 6 is concerned with simultaneously coordinating supplier selection and

202

Chapter 8

pricing and inventory decisions in a multi-level supply chain. The manufacturer

purchases optional components of certain functionality from his qualified suppliers to
produce a set of products for the retailers in different markets. The suppliers face the
problem to determine their optimal component prices sold to the manufacturer to
maximize the profits. The manufacturer makes decisions on the setup time interval for
production, the product wholesale prices, and the optimal suppliers and components
to maximize his profit. The retailers focus on the replenishment cycles and retail
prices of the products to achieve profit maximization. This problem is also modeled as
a three-level nested Nash game in which all the suppliers formulate the bottom-level
Nash game, the whole supplier sector play the middle-level Nash game with the
manufacturer, and both sectors as a group player formulate the top-level Nash game
with the retailers. Analytical and computational methods are developed to determine
the equilibrium of the game.
A numerical study in computer industry has been conducted to understand the
influence of market scale parameter, setup cost and component cost on the optimal
decisions and profits of the supply chain as well as its constituent members. In
addition, the influence of component selection strategy is also examined. There are
several research findings. When one retailers market scale increases, the sum of the
profits of the suppliers, the manufacturers profit and this retailers profit will increase,
while the other retailers profit will decrease. Furthermore, when the manufacturer
introduces a components selection strategy, the manufacturer tends to use high-end

203

Chapter 8

components for the product sold to this market. As a result, the manufacturer would
benefit from this strategy and the supply chain system includes fewer number of
suppliers and their total profit becomes lower. The manufacturer would setup his
production more frequently when the component selection strategy is employed.
8.1.5 CCSC with integration

Chapter 7 proposes an integrated approach for optimal supplier selection, pricing

and inventory decisions in a multi-level supply chain. The manufacturer obtains
components from the potential qualified suppliers and produces final products for
retailers in different markets. The suppliers, the manufacturer and the retailers would
like to be involved in an integrated manner to determine the pricing, inventory and
supplier selection decisions to maximize the overall supply chain system profit. We
formulate a mixed-integer programming model to integrate the optimal supplier
selection, pricing and inventory decisions. A cooperative game approach is employed
to evaluate the marginal contribution of the supply chain members. The integrated
supply chain model and cooperative game model are examined by a numerical study
and sensitivity analysis. The results show that: in an integrated manner, the rise of one
retail market scale increases the retail price and demand for the product in this market,
while without influencing on the other retail market; furthermore, an increased market
scale rises up the importance of this market and decreases the importance of the
suppliers provided components not used for the product in this market. We also show
that the change of setup cost has relatively little impact on the pricing decisions and
204

Chapter 8

marginal contributions of the supply chain members.

8.2 LIMITATIONS AND DIRECTIONS FOR FUTURE WORK

Despite achievements, this research, however, suffers from several limitations and
further extensions are necessary for future studies. We will summarize the limitations
and extensions in two parts. The first part includes the limitations and immediate
extensions of the five specific scenarios studied in this thesis. The second part
discusses the near or long term extensions including game models and supply chain
configuration models.
8.2.1 Limitation and immediate extensions for scenarios in this thesis
8.2.1.1 Scenario I
Our attention in this scenario has been focused on supply chains with one member
in each echelon. The competition at each echelon is also not covered. For example,
the suppliers may compete on pricing to provide materials to manufacturers. A more
general model with multiple suppliers, multiple manufacturers, and multiple retailers
could be developed.
The model in this scenario considers only one market strategic variable - price.
Carrying inventories is essential to enhance customer service and reduce distribution
costs. In the literature of channel competition, the impact of inventory on pricing
decisions has not been adequately considered. Thus, the involvement of inventory
factors into the supply chain issues is interesting and useful.
205

Chapter 8

In this scenario, we only consider two power structures for the decentralized and
semi-integrated channels (i.e., the Manufacturer-Stackelberg and the Nash games).
Some other power structures can be considered, such as the structure that the retailer
takes the channel leadership.
More sensitivity analyses could be conducted using more parameters and variables,
such as raw material cost, to understand the impacts of different power structures and
channel structures more thoroughly.
8.2.1.2 Scenario II
This scenario mainly focuses on coordination of the two-echelon supply chain. In
reality, a supply chain usually consists of multiple firms (suppliers, manufacturers,
retailers, etc.).
We assume that the production rate is greater than or equal to the demand rate to
avoid shortage cost. Without this assumption, the extra cost should be incorporated
into the future model.
The demand function employed by the scenario is linear demand function, which
is not enough to reflect the complicated retail markets. Some other demand functions,
such as iso-elastic, exponential demand functions, can be applied to this model to
analyze the optimal pricing and inventory decisions.
Besides, analysis and computational methods are incorporated to solve this model.
A more sophisticated solution algorithm is required to solve the problem for higher

206

Chapter 8

efficiency. Heuristic solution algorithm can be considered to solve this problem.

Finally, this scenario concentrates on one type of game structure. Some other ones
can be studied for other types of coordination relationships. For example, the
manufacturer can formulate Stackelberg game/ coalition game with the retailers. The
study and comparison on different game structures may provide guides for the
individual enterprises to choose their business policies and deserve further research
efforts.
8.2.1.3 Scenario III
Although this scenario considers multiple products and multiple retailers, the
competition among them is not covered. Under this competition, the demand of one
product / retailer is not only the function of his own price, but also the other products
/ retailers prices.
The suppliers are assumed to be selected and sole sourcing strategy is adopted. In
fact, either supplier selection or multiple sourcing is an inevitable part of supply chain
management.
Also, we assume that the production rate is greater than or equal to the demand
rate to avoid shortage cost. The assumption of production rate can be relaxed to
improve the generality.
Finally, this scenario focuses on one type of nested Nash game. Some other game
structures can be studied for other types of coordination relationships. For example,
the manufacturer can formulate a game with the retailers and then play games with his
207

Chapter 8

suppliers. Comparative study of different game structures might produce more

interesting managerial guidelines.
8.2.1.4 Scenario IV & Scenario V
For both scenarios, we assume that the components can be ranked in the order of
decreasing functionality, and either all or none the demand of a component is replaced
by the lower order component in the same SCS. These constraints can be relaxed and
the case that the components can be partially replaced by higher functionality
components.
The alternative suppliers are assumed to be qualified. We can extend to a two
phases model. We could consider the selection and qualification of a pool of suppliers
in the first stage, and then in the second stage, incorporating the qualified suppliers
into pricing and inventory decisions, which is the work done in Scenario IV.
The replenishment strategy between the suppliers and manufacturer is assumed to
be vendor managed inventory (VMI) strategy. Some other inventory system, such as
integer multipliers mechanism with each supply chain member holding their own
inventory, can be considered.
We can extend the situation that the production rate is limited. Thus, some extra
cost, such as the back-order penalty cost could be integrated into the model. The
competition among multiple products and among multiple retailers can also be
covered in the model extension as well.

208

Chapter 8

The locations of the supply chain members and the transportation modes as well
can be involved in the model extension.
Also, for Scenario IV, one type of nested Nash game is considered. Some other
game structures can be studied for other types of coordination relationships. For
example, the manufacturer can take the channel leadership and then formulate a game
with the retailers and the suppliers.
This Scenario V only puts forward a primary method of using cooperative game
theory to evaluate supply chain members in an integrated manner. How to connect this
supplier evaluation method with supplier selection decision deserves further
consideration. Besides, the Shapley value of the proposed cooperative game cannot be
regarded as a fair allocation rule of the total supply chain profit between the chain
members, because it might not be in the core. In the future work, we may focus on
working out fair allocation rule for the gains of the integrated supply chain.
8.2.2 Game models

This research considers several types of game models. However, most of them
focus on non-cooperative game approaches. It would be interesting to investigate
other types of game models to coordinate supply chain configuration decisions in
multi-echelon supply chain. The extension of the game models discussed in this
section falls into two categories: game structure extension which addresses the
different game structures that can be employed, such as sequential game, cooperative
game; and game composition extension, referring to combining different game
209

Chapter 8

structures into a whole game.

8.2.2.1 Game structure
The Nash game structures we mainly employed in this research figure out the
supply chain relationship without domination and of simultaneous decision-making
within the individual firms. In some situation, the firms of the supply chain might
have different market power and their action could not be simultaneously. Thus,
sequential game can be formulated. Yu et al. (2009) consider a Stackelberg game
between the manufacturer and its retailers of a two-echelon supply chain, where the
manufacturer takes the channel leadership and make the first move. The retailers, as
the follower, observe the manufacturers behavior and make their decisions. Extension
of such sequential game into multi-tier supply chain deserves further study.
Compared with non-cooperative game theory, the use of cooperative game is less
prevalent in supply chain management. Cooperative game directly considers possible
outcomes the individual firms can achieve. Nash bargaining model can be used to
determine supply chain configuration decisions and allocate of the profit pie between
the supply chain members. Coalition game is another widely studied cooperative
game, which focuses on formulates coalitions between the supply chain members and
divide the outcome, and whether the outcomes are stable. Both of them could be used
to analyze supply chain coordination problem in the further work.
8.2.2.2 Game composition
When the supply chain is complex, one game structure might not be enough to
210

Chapter 8

model the interaction between the chain members. There exist situations that
incorporate different game structures into an overall game of the supply chain
members for the different relationship between them. Take a three-echelon supply
chain (supplier-manufacturer-retailer) as an example, the manufacturer could
formulate Stackelberg game (or Nash bargaining game, etc.) with the suppliers and
then formulate Nash game (or Stackelberg game, etc.) with the retailers. Beside,
games can also be formulated within the suppliers and within the retailers. Obviously,
for multi-tier supply chain, the situation is much more complicated. Thus, various
types of game composition can be generated among the chain members. The
investigation of the different game composition can help model supply chain
coordination problems more accurately and provide guideline for the supply chain
members to choose their proper marketing and operation strategies.
8.2.3 Supply chain configuration

As we discussed in Chapter 1, supply chain configuration involves a large scale of

supply chain management problems. This research only focuses a few of them, such
as specific supply chain networks, pricing, inventory replenishment and supplier
selection decisions. This research could be extended along the aspects of supply chain
configuration addressed in Section 1.1: network design, sales and distribution,
logistics design, purchasing.
Firstly, this research mainly considers the supply chain with three-echelon and one
single manufacturer. The supply chains of multiple manufacturers are also rarely
211

Chapter 8

studied by researchers. The extension of the supply chain network to multiple

manufacturers case is more general and requires further investigation.
Ordering and marketing activities are important supply chain configuration
problems. Advertising, as an important marketing activity to attract customers,
especially for a new market, could be combined with pricing as a model extension.
From the logistics design view, the major inventory polices employed in this
research are multiple integer replenishment mechanism and vendor-managed
inventory strategies. Some other inventory policies, such as equal cycle time, integer
powers of two could be considered. It might provide interesting managerial
implications by comparing of these policies.
Purchasing decision in supply chain configuration concerns supplier selection and
material provided. However, other purchasing conditions, such as delivery price and
lead time are always determined or omitted. We can incorporate these elements into
further consideration.
Supply chain configuration decisions on warehouse locations and capacities,
inter-warehouse transportations, etc., in a supply chains distribution process also
have significantly impact on the final products delivery price, lead time (Chandra and
Grabis 2007). Thus, these decisions can also be involved in model extension.

8.3 OVERALL CONCLUSION

This thesis considers coordination of pricing, inventory and supplier selection

212

Chapter 8

decisions of multi-level supply chains using game theoretic approaches. Several novel
contributions have been made.
The first contribution is the formulation of game-theoretic models for five
different relevant scenarios of multi-level supply chain coordination problems as well
as the derivative of analytical proposition and implementation of the solution
algorithm. The model and solution procedures can serve as decision support tools for
the managers as a means of determining optimal configuration decisions and
understanding interaction between the supply chain members.
Secondly, we conduct comparative study between different power structure and
channel structures under pricing coordination in a multi-level supply chain with one
supplier, one manufacturer and one retailer. We prove that the integration for the
manufacturer and the retailer cannot always improve their profits in a monopoly for a
multi-level channel, which extends the conclusion of McGuire and Staelin (1983).
The third contribution is that we coordinate the pricing and inventory decisions for
two types supply chains with three-echelon inventory (i.e., one manufacturer and
multiple retailers; and multiple suppliers, one single manufacturer and multiple
retailers) through game-theoretic approach. Both of them integrate raw material
inventory into modeling. We also propose analytical and computational methods to
solve the game model. The proposed supply chain coordination models and solution
procedures are examined through some numerical examples. Some interesting
managerial implications can be derived from the experimental results.
213

Chapter 8

Finally, we incorporate supplier selection into pricing and inventory coordination

problem of multi-level supply chain. In order to demonstrate the applicability of the
proposed mathematical model and solution procedure, we consider a simulation case
from computer industry. The influences of different model parameters and component
selection strategy are studied. We derive some interesting observations from the
experimental results.

214

APPENDIX
Appendix A.

b 1
b 1
b 2b b 3
3b 1

Proof.

is a decrement function of b.

We can write:

b 1
b

2b

3b 1

b 3

b 1

ln

b 13b1
b1
b b 3
2b

3b 1 ln b 1 2 b ln b b 1 ln b 3

3b 1 ln b 1 2b ln b b 1 ln b 3

Thus, we just need to prove

is

decrement function.
Denote x b 3b 1 ln b 1 2b ln b b 1 ln b 3 .
Take the first derivative, we have:

x b 3ln b 1 2ln b ln b 3
'

This is because x b
''

2
2

0.
b 1 b 3

2 3b3 11b 2 13b 9

b b 1 b 3
2

0.

When b tends to infinite, x b tends to the highest value. That is:

'

b 13
2
2
lim x b lim ln 2

0.
b
b

b
b

3
b

1
b

'

For any b (b>3), x b <0, thus, x b is a decrement function.

'

b 1
b 1
2b
b b 3
3b 1

Hence,

is a decrement function with b.

Appendix B.
Proof of Proposition 3.1(b). Compare the retail price and the raw material price in
the MR-S case with those in the MR-N case:

215

Appendix

prMR S b b 2
b2 2b

1
prMR N b 12 b 2 2b 1
psMR S
s cs cm
1
0
2
MR N
ps
b 1 b 1 s cs cm
Hence, prMRS prMR N , psMRS psMR N . That is the equilibrium prices for chain
members in the MR-S case are no higher than those in the MR-N case.
Compare the retailer and the manufacturers profit, the suppliers profit and the
entire system profit in the MR-S case with those in MR-N case:
MR S
b 1
mr
b
1.
b 1
MR N
mr
b b 2
2 b 1

It is decrement function of b. When b tends to infinite,

MR S
mr
tends to the
MR N
mr

lowest value 1.

sMR S b 1
Similarly, we have: MR N

b b 2
s

b 1

MR S
1 . So MR N 1 .

The profits for the chain members and the entire supply chain system in the MR-S
MR S
MR N
case are no less than those in the MR-N case: mr
, sMR S sMR N ,
mr

MRS MR N . This completes the proof of part(b).

216

REFERENCE
Abad, P.L., 1988, Determining optimal selling price and lot size when the supplier
offers all-unit quantity discounts, Decision Sciences, 19, 622634.
Abad, P. L., 1994, Supplier pricing and lot sizing when demand is price sensitive,
European Journal of Operational Research, 78, 334354.
Abad, P.L., 1997, Optimal policy for a reseller when the supplier offers a temporary
reduction in price, Decision Sciences, 28 (3), 637653.
Abad, P.L., Jaggi, C.K., 2003, A joint approach for setting unit price and the length of
the credit period for a seller when end demand is price sensitive, International
Journal of Production Economics, 83, 115122.
Alan,

G.K.,
Medini
R.
Singh,
1992,
Pricing
in
a
manufacturer-retailer-customer system, technical report, 48-92.

three-tier

Alchian, A, Demsetz, H., 1972, Production, information costs, and economic

organization. American Economic Review, 62, 777-795.
Arcelus, F.J., G. Srinivasan, 1987, Inventory policies under various optimizing
criteria and variable markup rate, Management Science, 33(6), 756-762.
Arcelus1, F.J. , Srinivasan1, G., 1998, Ordering policies under one time only discount
and price sensitive demand, IIE Transactions, 30, 1057-1064.
Ardalan, A., 1991, Combined optimal price and optimal inventory replenishment
policies when a sale results in increase in demand, Computers and Operations
Research, 18(9), 721-730.
Ardalan A., 1994, Optimal prices and order quantities when temporary price discounts
result in increase in demand, European Journal of Operational Research, 72,
5261.
Ardalan, A., 1995, A comparative analysis of approaches for determining optimal
price and order quantity when a sale increases demand, European Journal of
Operational Research, 84(2), 416-430.
Arntzen, B.C., Brown, G.G., Harrison, T.P., Trafton, L.L., 1995, Global Supply Chain
Management at Digital Equipment Corporation, Interfaces, 25, 69-93.
Aumann, R.J., 1959, Acceptable points in general cooperative N-person game, In
A.W. Tucker and R.D. Luce, editors, Contributions to the Theory of Games,
4, Princeton University Press, Princeton, NJ.
Banerjee, A., 1986, A joint economic-lot-size model for purchase and vendor,
Decision Sciences, 17, 292311.
217

Reference

Banerjee, A., Banerjee, S., 1992, Coordinated, orderless inventory replenishment for a
single supplier and multiple buyers through electronic data interchange,
International Journal of Technology Management, 7, 328336.
Bergen, M., John, G., 1997, Understanding cooperative advertising participation rates
in conventional channels, Journal of Marketing Research, 34(3), 357-369.
Bernstein F., Federgruen A., 2003, Pricing and replenishment strategies in a
distribution system with competing retailers, Operations Research, 51(3),
409-426.
Boyaci, T., Gallego, G., 2002, Coordinating pricing and inventory replenishment
policies for one wholesaler and one or more geographically dispersed retailers,
International Journal of Production Economics, 77, 95-111.
Brookes Richard, 1995, Recent changes in the retailing of fresh produce: Strategic
implications for fresh produce suppliers, Journal of Business Research, 32(2),
149-161.
Burns Robbie, 1994, Capability is where the action is, Logistics Information
Management, 7(5), 58-60.
Burwell, T.H., Dave, D.S., Fitzpatrick, K.E., Roy, M.R., 1990, A note on determining
optimal selling price and lot size under all-unit quantity discounts, Decision
Sciences, 21(2), 471- 474.
Byrne, M.D., Bakir, M.A., 1999, Production Planning Using a Hybrid Simulation
Analytical Approach, International Journal of Production Economics, 59,
305-311.
Buzzell, R.D., Quelch, J.A., Salmon, W.J., 1990, The costly bargain of trade
promotion, Harvard Business Review, 141149.
Cachon, G.P., Netessine, S., 2004, Game theory in supply chain analysis, In: D.
Simchi-Levi, S.D. Wu and M. Shen, Editors, Supply Chain Analysis in the
eBusiness Era, Kluwer.
Chakravarty, A., Martin, G.E., 1988, An optimal joint buyer-seller discount pricing
model, Computers and Operations Research, 15(3), 271-281.
Chan, LMA, ZJM Shen, D Simchi-Levi, JL Swann, 2004, Coordination of pricing and
inventory decisions: A survey and classification, International Series in
Operations Research and Management, Chapter 14.
Chandra, C., Grabis, J., 2007, Supply Chain Configuration Concepts, Solutions, and
Applications, Springer, New York.
Chen, C.K., Min, K.J., 1994, An analysis of optimal inventory and pricing policies
under linear demand, Asia-Pacific Journal of Operational Research, 11(2),
117-129.
218

Reference

Chen, F.Y., H Yan, L Yao, 2004, A newsvendor pricing game, IEEE Transactions on
System, Man, and Cybernetics, 34(4), 450-456.
Chen Miao-Sheng, Horng-Jinh Chang, Chih-Wen Huang, Chin-Nung Liao, 2006,
Channel coordination and transaction cost: A game-theoretic analysis,
Industrial Marketing Management, 35, 178-190.
Cheng T.C.E., 1990, An EOQ model with pricing consideration, Computers and
Industrial Engineering, 18(4), 529-534.
Cheng, F., Sethi., S.P., 1999, A periodic review inventory model with demand
influenced by promotion decisions, Management Science, 45(11), 1510-1523.
Chiang, W., Fitzsimmons, J., Huang, Z., Li, S.X., 1994, A game-theoretic approach to
quantity discounts problems. Decision Science, 25, 153-168.
Chintagunta, P.K., 2002, Investigating category pricing behavior at a retail chain,
Journal of Marketing Research, 39(2), 141-154.
Choi, S.C., 1991, Price competition in a channel structure with a common retailer,
Marketing Science, 10(4), 271-296.
Choi, S.C., 1996, Price competition in a duopoly common retailer channel, Journal of
retailing, 72(2), 117-134.
Chopra, S., Meindl, P., 2001, Supply chain management: strategy, planning and
operations, Upper Saddle River: Prentice Hall.
Cohen, M.A., Lee, H.L., 1988, Strategic Analysis of Integrated Production
Distribution System: Models and Methods, Operations Research, 36(2),
216-228.
Cohen, M.A., Lee, H.L., 1989, Resource Deployment Analysis of Global
Manufacturing and Distribution Networks. Journal of Manufacturing
Operations Management, 2, 81-104.
Corbett, C., de Groote, X., 2000, A suppliers optimal quantity discount policy under
asymmetric information, Management Science, 46(3), 444-450.
Coyle, J.J., E.J. Bardi, C.J. Langley, 1996, A management guide to cooperative
advertising, Association of National Advertisers, New York.
Curran, T.A, A. Ladd, 2000, SAP/R3 business blueprint: understanding enterprises
supply chain management, Upper Saddle River: Prentice Hall.
Dada, M., Srikanth K.N., 1987, Pricing policies for quantity discounts, Management
Science, 33, 1247-1252.
de Boer, L., Labro, E., Morlacchi, P., 2001. A review of methods supporting supplier
selection, European Journal of Purchasing & Supply Management, 7, 7589.
219

Reference

Degraeve, Z., F. Roodhooft, 1999, Improving the efficiency of purchasing process

using total cost of ownership information: the case of heating electrodes at
cockerill sambre s.a., European Journal of Operations Research, 112, 42-53.
Degraeve, Z., Roodhooft, F., 2000, A mathematical programming approach for
procurement using activity based costing, Journal of Business, Finance and
Account, 27 (1/2), 6998.
De Toni, A.F., Zamolo, E., 2005, From a traditional replenishment system to
vendor-managed inventory: A case study from the household electrical
appliances sector, International Journal of Production Economics, 96(1),
6379.
Disney, S.M., Towill, D.R., Bullwhip reduction in supply chains: The impact of VMI,
International Journal of Operations and Production Management, 23(6),
625651.
Dominguez, H., Lashkari, R.S., 2004, Model for integrating the supply chain of an
appliance company: a value of information approach, International Journal of
Production Research, 42(11), 2113-2140.
Dong, Y., Xu, K.F., 2002, A supply chain model of vendor managed inventory,
Transportation Research Part E Logistics and Transportation Review, 38(2),
7595.
Driessen, T., 1988, Cooperative game, solutions and applications, Kluwer Academic
Publishers.
Edward C. Rosenthal, 2008, A game-theoretic approach to transfer pricing in
vertically integrated supply chain, International Journal of Production
Economics, 115, 542-552.
Elmaghraby W., Keskinocak P., 2003, Dynamic pricing in the presence of inventory
considerations: research overview, current practices, and future directions,
Management Science, 49(10), 1287-1309.
Esmaeili, M., Mir-Bahador Aryanezhad, P. Zeephongsekul, 2009, A game theory
approach in seller-buyer supply chain, European Journal of Operations
Research, 191(2), 442-448.
Federgruen, A., A. Heching, 2002, Multilocation combined pricing and inventory
control, Manufacturing and Service Operations Management, 4(4), 275-295.
Fisher ML, 2003, What is the right supply chain for your product? Operations
Management: Critical Perspectives on Business and management, Harvard
business review, March-April 1997, 105-160.
Foley, S., Mahmood, T., 1996, Wal-Mart store, Inc. Harvard Business Review
(August).

220

Reference

Geoffrion, A.M., Graves, G.W., 1974, Multi-commodity Distribution System Design

by Benders, Decomposition, Management Science, 20, 822-844.
Gerchak, Y., D., Gupta, 1991, On apportioning costs to customers in centralized
continuous review systems, Journal of Operations Management, 10, 546-551.
Gerstner, E., Hess, J.D., 1995, Pull promotion and channel coordination, Marketing
Science, 14(1), 43-60.
Ghodsypour, S., OBrien, C., 2001, The total cost of logistics in supplier selection,
under conditions of multiple sourcing, multiple criteria and capacity constraint,
International Journal of Production Economics, 73, 15-27.
Giannoccaro, L., Pontrandolfo, P., 2004, Supply chain coordination by revenue
sharing contracts, International Journal of Production Economics, 89(2),
131-139.
Goyal, S.K., 1977, An integrated inventory model for a single supplier-single
customer problem, International Journal of Production Research, 15(1),
107-111.
Goyal, S.K., 1988, A joint economic-lot-size model for purchase and vendor: A
comment, Decision Sciences, 19, 236-241.
Goyal, S.K., 1995, A one-vendor multi-buyer integrated inventory model: A comment,
European Journal of Operational Research, 82(1), 209-210.
Granot, D., G. Soi, 2005, Formation of alliances in internet-based supply exchanges,
Management Science, 51, 92-105.
Graves, S.C., Willems, S.P., 2005, Optimizing the Supply Chain Configuration for
New Products, Management Science, 51, 1165-1180.
Hall, R., 1992, Price changes and order quantities: impacts of discount rate and
storage costs, IIE Transactions, 24(2), 104110.
Hamedi, M., Reza Z.F., Mohammad M.H., Gholam R.E., 2009, A distribution
planning model for natural gas supply chain: A case study, Energy Policy,
37(3), 799-812.
Haq, A.N., Kannan, G., 2006, Design of an integrated supplier selection and
multi-echelon distribution inventory model in a built-to-order supply chain
environment, 44, 1963-1985.
Hartman, B., Dror, M., Shaked, M., 2000, Cores of inventory centralization games,
Games and Economic Behavior, 31, 26-49.
Hill, R.M., 1997, The single-vendor single-buyer integrated production-inventory
model with a generalized policy, European Journal of Operational Research,
97(3), 493-499.
221

Reference

Hoch, S.J., Kim, B., Alan L.M., Peter E.R., 1995, Determinants of store-level price
elasticity, Journal of Marketing Research, 32(1), 17-29.
Hopp, W., Spearman, M., 1993, Setting safety lead times for purchased components
in an assembly system, IIE Transactions, 25, 2-11.
Huang, Z., Li, S.X., Mahajan, V., 2002, An analysis of manufacturer-retailer supply
chain coordination in cooperative advertising, Decision Science, 33(3),
469-494.
Huang, G.Q., Zhang, X.Y., Liang, L., 2005b, Towards integrated optimal
configuration of platform products, manufacturing processes, and supply
chains, Journal of Operations Management, 23, 267-290.
Huang, G.Q., Zhang, X.Y., Lo, V.H.Y., 2007, Integrated configuration of platform
products and supply chains for mass customization: a game-theoretic approach,
IEEE Transactions on Engineering Management, 54(1), 156-171.
Ingene, C.A., Parry, M.E., 1995a, Channel coordination when retailers compete,
Marketing Science 14 (4), 360-377.
Ingene, C.A., Parry, M.E., 1995b, Coordination and manufacturer profits
maximization: The multiple retailer channel, Journal of Retailing 72 (2),
139-151.
Ingene, C.A., Parry, M.E., 1998, Manufacturer-optimal wholesale pricing when
retailers compete, Marketing Letters, 9 (1), 65-77.
Ingene, C.A., Parry, M.E., 2000, Is channel coordination all it is cracked up to be?
Journal of Retailing, 76 (4), 511547.
Jabber, M.Y., Goyal, S.K., 2008, Coordinating a three-level supply chain with
multiple suppliers, a vendor and multiple buyer, International Journal of
Production Economics, 11(6), 95-103.
Jensen, M, Meckling, W., 1976, Theory of the firm: managerial behavior, agency
costs and ownership structure. Journal of Financial Economics, 3, 305-360.
Jeuland A. P., Shugan S. M., 1983, Managing channel profits, Marketing Science, 2,
239-272.
Jeuland, A.P., Shugan, S.M., 1988, Channel of distribution profits when channel
members form conjectures, Marketing Science, 7(2), 202-210.
Joseph G.S., Zhang, Jiang 2009, Pricing and two-tier advertising with one
manufacturer and one retailer, European Journal of Operational Research,
192(3), 904-917.
Kachani, S., Perakis, G., 2002, A fluid model of dynamic pricing and inventory
management for make-to-stock manufacturing systems, Working Paper, MIT.
222

Reference

Karray, S., Zaccour, G., 2006, Could co-op advertising be a manufacturers

counterstrategy to store brands? Journal of Business Research, 59, 1008-1015.
Kim, B.J., Leung, M.Y., Park, K.T., Zhang, G.O., Lee, S., 2002, Configuring a
Manufacturing Firm's Supply Network with Multiple Suppliers, IIE
Transactions, 34, 663-677.
Kim, Daesoo, Lee, W.J., 1998, Optimal joint pricing and lot sizing with fixed and
variable capacity, European Journal of Operational Research, 109 (1),
212227.
Kim, Jin-Ho, Park, Jong-Bae, Park, Jong-Keun, Kim, B H, 2003, A new
game-theoretic framework for maintenance strategy analysis, IEEE
Transactions on Power Systems, 18, (2), 698-706.
Kim, K.H., Hwang, H.,1989, Simultaneous improvement of suppliers profit and
buyers cost by utilizing quantity discount, Journal of the Operational
Research Society, 40, 255-265.
Ko, H.J., Ko, C.S., Kim, T., 2006, A Hybrid Optimization/Simulation Approach for a
Distribution Network Design of 3PLS, Computers and Industrial Engineering,
50(4), 440-449.
Kohli, R., Park, H., 1989, A cooperative game theory model of quantity discounts,
Management Science, 35(6), 693-707.
Kuhn, H.W., 1953, Extensive games and the problem of information, In H.W. Kuhn
and A.W. Tucker, editors, Contributions to the Theory of Games, 2,
Princeton University Press, Princeton, NJ.
Kunreuther, H., J. F. Richard, 1971, Optimal pricing and inventory decisions for
nonseasonal item. Econometri ca, 39(1), 173-175.
Lal, R., Staelin, R., 1984, An approach for developing an optimal discount pricing
policy, Management Science, 30, 1524-1539.
Lariviere, M.A., Porteus, E.L., 2001, Selling to the newsvendor: an analysis of
price-only contracts, Manufacturing and Service Operations Management, 3,
293-305.
Lau, Amy Hing Ling, Lau Hon-Shiang, 2003, Effects of a demand-curves shape on
the optimal solutions of a multi-echelon inventory / pricing model, European
Journal of Operational Research, 147, 530-548.
Lee, H.L., Billington, C., 1995, The evolution of supply chain management models
and practice at Hewlett-Packard, Interfaces, 25(5), 42-63.
Lee, Y.H., Kim, S.H., 2000, Optimal production-distribution planning in supply chain
management using a hybrid simulation-analytic approach, Proceedings of the
2000 Winter Simulation Conference, 1252-1259.
223

Reference

Lee, Y.H., Kim, S.H., 2002, Production-distribution planning in supply chain

considering capacity constraints, Computers and Industrial Engineering, 43,
169-190.
Lee, H.L., Rosenblatt, J., 1986, A generalized quantity discount pricing model to
increase suppliers profits, Management Science, 33, 1167-1185.
Lee, W.J., Kim, Daesoo, 1993, Optimal and heuristic decision strategies for integrated
product and marketing planning, Decision Sciences, 24 (6), 1203-1213.
Lee, W.J., 1993, Determining order quantity and selling price by geometric
programming: Optimal solution, bounds and sensitivity, Decision Sciences, 24,
76-87.
Leng, M., M., Parlar, 2009, Allocation of cost savings in a three-level supply chain
with demand information sharing: a cooperative-game approach, Operations
Research, 57, 200-213.
Li, S.X., Huang, Z., Ashley, A., 1996, Inventory, channel coordination and bargaining
in a manufacturerretailer system, Annals of Operations Research, 68, 4760.
Liao Zhiying, J. Rittscher, 2007, Integration of supplier selection, procurement lot
sizing and carrier selection under dynamic demand condition, 107, 502-510.
Liu, Baoding, 1998, Stackelberg-Nash equilibrium for multilevel programming with
multiple followers using genetic algorithms, Computer Math. Application,
36(7), 79-89.
Liu, L., Parlar, M., Zhu, S.X., 2007, Pricing and lead time decisions in decentralized
supply chains, Management Science, 53(5), 713725.
Liu Liming, M. Parlar, S.X. Zhu, 2007, Pricing and lead time decisions in
decentralized supply chain, Management Science, 53(5), 713-725.
Lu, L., 1995, A one-vendor multi-buyer integrated inventory model, European Journal
of Operational Research, 81, 312-323.
Martha C. Cooper, Lisa M. Ellram, 1993, Characteristics of supply chain management
and the implications for purchasing and logistics strategy, The International
Journal of Logistics Management, 4, 13-24.
Martin, G.E., 1994, Note on an EOQ model with a temporary sale price. International
Journal of Production Economics, 37(2-3), 241-243.
McGuire, T.W., Richard Staelin, 1983, An industry equilibrium analysis of
downstream vertical integration, Marketing Science, 2, 161-190.
McIvor, R., Humphreys, P., 2004, Early supplier involvement in the design process:
lessons from the electronics industry, Omega, 32(3), 179-199.

224

Reference

Mc Williams, 1997, Whirlwind on the web, Business Week, April 7, 132-136.

Meca A., 2000, Juegos cooperativos asociados a problemas de inventario. PhD
dissertation, Miguel Hernandez University
Meca A., Timmer J., Garcia-Jurado I., Borm P., 1999, Cooperation and competition in
inventory games, Mathematical Methods of Operations Research, 57, 481-493.
Meca, A., Timmer, J., Garcia-Jurado, I., Borm, P., 2004, Inventory Games, European
Journal of Operational Research, Vol. 156, pp. 127-139.
Michael, S.C., 2002, Can a franchise chain coordinate?, Journal of Business
Venturing, 17(4), 325-341.
Michael Tracey, Chong Leng Tan, 2001, Empirical analysis of supplier selection and
involvement, customer satisfaction, and firm performance, Supply Chain
Management: An International Journal, 6(4), 174-188.
Mihaly, M., 1999, Strategic sourcing, Industry Week, 248(4), 3-8.
Minakshi Trivedi, 1998, Distrubution channels: an extension of exclusive relationship,
Management Science, 44(7), 896-909.
Morgan, J., 1996, Integrated supply chains: how to make them work, Purchasing,
120(1), 110-113.
Monahan, J.P., 1984, A quantity discount pricing model to increase vendor profits,
Management Science, 30(6), 720-726.
Monczka, R.M., Ragartz, G.L., Handfield, R.B., 1997, Supplier integration into new
product development: preliminary results, Advances in the Management of
Organizational Quality, 2, 87-138.
Moorthy, K.S., 1987, Managing channel profits: comment, Management Science, 6(4),
375-379.
Moorthy, K.S., 1988, Strategic Decentralization in Channels, Marketing Science, 7,
335-355.
Morgan, R. L. Daniels, P. Kouvelis., 2001 Marketing/manufacturing trade-offs in
product line management, IIE Transactions, 33(11), 949-962.
Moutaz Khouja, 2003, Optimizing inventory decisions in a multi-stage
multi-customer supply chain, Transportation Research Part E, 39, 193-208.
Munson, C.L., Rosenblatt, M.J., 1997, The impact of local content rules on global
sourcing decisions. Production and Operations Management, 6, 277-289.
Munson, C.L., Rosenblatt, M.J., 2001, Coordinating a three-level supply chain with
quantity discounts, IIE Transactions, 33, 371-384.
225

Reference

Nagarajan M., Soi G.., 2006, Game-theoretic analysis of cooperation among supply
chain agents: review and extensions, European Journal of Operations Research,
187(3), 719-745.
Nash, J.F., 1950, Equilibrium points in N-person game, Proceedings of the National
Academy of Sciences of the United States of America, 36(1), 48-49.
Oliver R.K., Michael D. Webber, 1992, Supply chain management: logistic catches
up with strategy, Outlook 1982, cit. Martin Christopher, Logistics, The
Strategic Issues, London: Chapman and Hall.
Osborne, M.J., Rubinstein, A., 1994, A course in game theory, Massachusetts Institute
of Technology Press, Massachusetts.
Park, Myung-Ho, Mason, J. Barry, 1990, Toward an integrated model of determinants
of business performance, Research in Marketing, 10, 157-201.
Parlar, M. and Wang, Q., 1994, Discounting decisions in a supplier buyer relationship
with a linear buyers demand, IIE Transactions, 26(2), 3441.
Periaux, J., H.Q. Chen, B. Mantel, M. Sefrioui, H.T. Sui, 2001, Combining game
theory and genetic algorithms with application to DDM-nozzle optimization
problems, Finite Elements in Analysis and Design, 37, 417-429.
Peters Brett A., Rajasekharan Maheswaran, 1996, A genetic algorithm for
determining facility design and configuration of single-stage flexible
electronic assembly system, Journal of Manufacturing Systems, 5(15),
316-324.
Poirier, C.C., 1999, The convergence of business and technology, Supply Chain
Management Review, 3(3), 52-58.
Porter, M., 1985, Competitive advantage: creating and sustaining superior
performance, New York: The Free Press.
Prafulla Joglekar, Madjid Tavana, Jack Rappaport,2006, A comprehensive set of
models of intra and inter-organizational coordination for marketing and
inventory decisions in a supply chain, International Journal of Integrated
Supply Management.
Ragatz, G.L., Handfield, R.B., Scannell, T.V., 1997, Success factors for integrating
suppliers into new product development, Journal of Production Innovation
Management, 14, 190-202.
Rangan, V. Kasturi, Melvyn A. J. Menezes, E. P. Maier, 1992, Channel selection for
new industrial products: a framework, method, and application, Journal of
Marketing, 56(3), 69-82.
Rhim, H., Ho, T.H., Karmarkar, U.S., 2003, Competitive location, production, and
market selection, European Journal of Operational Research, 149, 211-228.
226

Reference

Robert B. Handfield, Gary L. Ragatz, Kenneth J. Petersen, Robert M. Monczka, 1999,

Involving suppliers in new product development, California Management,
Review, 42(1), 59-82.
Romano 2003 Romano, P. (2003), Co-ordination and integration mechanisms to
manage logistics processes across supply networks, Journal of Purchasing and
Supply Management, 9(3), 119-134.
Rusdiansyah, A., Tsao, D.B., 2005, An integrated model of the periodic delivery
problems for vending-machine supply chains, Journal of Food Engineering,
70(3), 421-434.
Samuelson, P. A., 1947, Foundations of economic analysis. Cambridge, Harvard
University Press
Sandoe, K., G. Corbitt, R. Boykin, 2001, Enterprise integration. New York: John
Wiley and Sons.
Shapiro, J.F., 2001, Modeling the Supply Chain, Duxbury Press.
Shapley, L.S., 1971, Cores of convex games, International Journal of Game Theory,
1(1), 11-26.
Sarker, B. R., Parija, G.R., 1994, An optimal batch size for a production system
operating under a fixed-quantity, periodic delivery policy, Journal of the
Operational Research Society, 45 (8), 891-900.
Shugan, Steven M., 1985, Implicit understandings in channels of distribution,
Management Science, 31(4), 435-460.
Simchi-Levi, D., P. Kaminsky, E. Simchi-Livi, 2000, Designing and Managing the
Supply Chain-Concepts, Strategies and Case Studies, McGraw-Hill,
Singapore.
Slikker, M., Fransoo, J., Wouters, M., 2005, Cooperation between multiple
news-vendors with transshipments, European Journal of Operational Research,
167(2), 370-380.
Smirnov, A.V., Sheremetov, L.B., Chilov, N., Cortes, J.R., 2004, Soft-computing
technologies for configuration of cooperative supply chain, Applied Soft
Computing, 4(1), 87-107.
Syarif, A., Yun, Y., Gen, M., 2002, Study on multi-stage logistic chain network-a
spanning tree-based genetic algorithm approach, Computers and Industrial
Engineering, 43, 299-314.
Talluri, S., Baker, R.C., 2002, A multi-phase mathematical programming approach for
effective supply chain design, European Journal of Operational Research, 141,
544-558.

227

Reference

Tan, K. C., 2001, A framework of supply chain management literature, European

Journal of Purchasing & Supply Management, 7, 39-48.
Tavakkoli-Moghaddam, R., A.R. Rahimi-Vahed, Multi-criteria sequencing problem
for a mixed-model assembly line in a JIT production system, Applied
Mathematics and Computation, 181, 1471-1481.
Tersine, R.J., Price, R.L., 1981, Temporary price discounts and EOQ, Journal of
Purchasing and Materials Management, 17(4), 23-27.
Thomas, D.J., Griffin, P.M., 1996, Coordinated supply chain management, European
Journal of Operational Research, 94, 1-15.
Truong, T.H., Azadivar, F., 2005, Optimal Design Methodologies for Configuration
of Supply Chains, International Journal of Production Research, 43(11),
2217-2236.
Tullous, R., Utrecht, RL, 1992, Multiple or single sourcing, Journal of Business and
Industrial Marketing, 7(3), 5-18.
Tyan, J., Wee, HM, 2003, Vendor managed inventory: A survey of the Taiwanese
grocery industry. Journal of Purchasing and Supply Management, 9(1), 11-18.
Venkatraman, N., John E. Prescott, 1990, Environment-strategy coalignment: an
empirical test of its performance, Implications Management Journal, 11, 1-24.
Viswanathan S., Wang, Q., 2003, Discount pricing decisions in distribution channels
with price-sensitive demand, European Journal of Operational Research, 149,
571-587.
Von Neumann J., O. Morgenstern, 1944, Theory of games and economic behavior,
Princeton University Press, Princeton, NJ.
Wang, Q., 2002, Determination of suppliers optimal quantity discount schedules with
heterogeneous buyers, Naval Research Logistics, 49, 46-59.
Wang, Y., Gerchak, Y., 2003, Capacity games in assembly systems with uncertain
demand, Manufacturing and Service Operations Management, 5(3), 252-267.
Watts, C., Hahn, C., 1993, Supplier development programs: an empirical analysis.
International Journal of Purchasing and Materials Management, 29, 1017.
Weber, C., Current, J., 1993, A multi-objective approach to vendor selection,
European. Journal of Operations Research, 68, 173184.
Weng, ZK, Wong, RT, 1993, General models for the suppliers all-unit quantity
discount policy, Naval Research Logistics, 40(7), 971-991.

228

Reference

Weng, Z.K, 1995a, Modeling quantity discounts under general price-sensitive demand
functions: optimal policies and relationships, European Journal of Operational
Research, 86, 300314.
Weng, Z.K, 1995b, Channel coordination and quantity discounts, Management
Science, 41(9), 1509-1522.
Whitin, TM, 1955, Inventory control and price theory, Management Science, 2(1),
61-68.
Williamson, O., 1975, Markets and hierarchies: analysis and antitrust implications,
The Free Press, New York.
Woo Y. Y., Shu-Lu Hsu, Soushan Wu, 2001, An integrated inventory model for a
single vendor and multiple buyers with ordering cost reduction, 73, 203-215.
Wu, T., OGrady, P., 2001, A network based approach to the design of supply chains,
Internet Lab Technical Report, Department of Industrial Engineering,
Seamans Center, University of Iowa.
Yang, P.C., Wee, H.M., 2003, Optimal strategy in vendor-buyer alliances with
quantity discount, International Journal of Computer Integrated Manufacturing,
16(6), 455-463.
Yano, C., 1987a, Setting planned lead times in serial production systems with
tardiness costs, Management Science, 33, 95-106.
Yano, C., 1987b, Stochastic lead times in two-level distribution systems, Naval
Research Logistic, 34, 831-843.
Yano, C., 1987c, Stochastic lead times in two-level assembly systems, IIE
Transactions, 19, 371-378.
Yao, Y.L., P.T. Evers, M.E. Dresner, 2007, Supply chain integration in
vendor-managed inventory, Decision Support Systems, 43(2), 663-674.
Yu, Yugang, Feng Chu, Haoxun Chen, 2009a, A Stackelberg game and its
improvement in a VMI system with a manufacturing vendor, European
Journal of Operational Research, 192(3), 929-948.
Yu, Yugang, George Q. Huang, Liang Liang, 2009b, Stackelberg game-theoretic
model for optimizing advertising, pricing and inventory policies in vendor
managed inventory (VMI) production supply chains, Computers and Industrial
Engineering, 57(1), 368-382.
Yue, J., J. Austin, M. Wang, Z. Huang, 2006, Coordination of cooperative advertising
in a two-level supply chain when manufacturer offers discount, European
Journal of Operational Research, 168, 65-85.

229

Reference

Zhu, K., U. W. Thonemann, 2002, Coordination of pricing and inventory control

across products. Working Paper, Stanford University.

230

PUBLICATIONS DURING PH.D. STUDY

Publications in Journals
Invited Book Chapter
Yun Huang and George Q. Huang, Nash Game-Theoretic Model for Optimizing
Pricing and Inventory Policies in a Three-Level Supply Chain. Accepted for
publication in IAENG Electrical Engineering and Applied Computing, Special
Edition of the World Congress on Engineering 2010.
George Q. Huang and Yun Huang, 2009, Game-theoretic coordination of pricing and
inventory decisions in a multi-level supply chain, International Journal of
Production Economics, Major revision.
Yun Huang and George Q. Huang, 2010, Coordinated selection of suppliers and
components considering pricing and inventory in a multi-level supply chain: a
game-theoretic approach, Transportation Research Part E: Logistics and
Transportation Review, Accepted.
Yun Huang and George Q. Huang, 2010, Price coordination in a three-level supply
chain with different channel structures using game-theoretic approach,
International Journal of Management Science and Engineering Management, 5,
83-94.
Yun Huang and George Q. Huang, 2010, Joint Pricing and Inventory Replenishment
Decisions in a Multi-level Supply Chain, Accepted for special issue of IAENG
Journal.
Yun Huang and George Q. Huang, 2010, Game theoretic coordination of marketing
and inventory policies in a two-level supply chain with competing retailers,
International Journal of Logistics, Submitted.
Yun Huang and George Q. Huang, 2010, Simultaneous coordination of supplier
selection, pricing and replenishment decisions in a Multi-level Supply Chain,
Computer and Industrial Engineering, Submitted.
Publications in Conferences
Yun Huang and George Q. Huang, 2010, Game-theoretic coordination of marketing
and inventory policies in a multi-level supply chain, Proceedings of the World
Congress on Engineering (Best student paper award).
Yun Huang and George Q. Huang, 2009, Game theory approach for coordinating
marketing and inventory policies in a two-level supply chain with retailers
competition, Proceedings of INFORMS International Conference on Service
Science.

231

Yun Huang and George Q. Huang, 2008, Optimal pricing for a three-level supply
chain with different channel structures using game-theoretic approach,
Proceedings of Hamburg International Conference of Logistics.

232