Beruflich Dokumente
Kultur Dokumente
Advisor(s)
Huang, GQ
Author(s)
Huang, Yun;
Citation
Issued Date
URL
Rights
2010
http://hdl.handle.net/10722/130938
By
Yun Huang
B.Sc. SEU
M.Sc. ZJU
June 2010
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
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
iv
vi
LIST OF FIGURES
vii
LIST OF TABLES
viii
LIST OF SYMBOLS
market scale
A constant in the demand function of retailer l, which represents his market
Arl
scale
b
cLij
cm
cml
cslv
csv
Drl
erll
erlj
product l
FLij
Gml
Grl
Gsv
hmpl
hmrs
hmslv
hrl
hsv
krl
K sv
v
L
mm
List of symbols
mr
nml
Oml
Ordering processing cost of raw materials used for product l per order
Orl
Osv
Pl
pLij
Decision variable, the price of component Lij charged by the supplier to the
manufacturer
pm
pml
pr
prl
ps
psv
Qsv
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
slvl
Sm
sv
tsv Lij
uLik l
List of symbols
Vl
wsv
Xm
X rl
X sv
z Lij
Zm
Z rl
Z sv
sv
ijk
Binary decision variable to indicate whether component Lij has been used
lvl
to replace Lik
xi
CHAPTER 1
INTRODUCTION
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
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
Pricing
Inventory
Supplier
selection
p
up
...
Cooperative game
Non-cooperative game
ch
ly
Advertising
o
tw
ne
n
ai
rk
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
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.
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
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
Chapter 1
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
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
18
Chapter 1
Chapter 1
20
Chapter 1
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,
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
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
27
Chapter 2
28
Chapter 2
Characteristic
Reference
Whitin 1955; Kunreuther and Richard 1971;
Horizontal
Incorporating suppliers pricing
strategy
One supplier and one buyer (or
homogeneous buyers)
Vertical
Gallego 2002
Cheng 1990; Chen and Min 1994b; Morgan et al.
2001; Kachani and Perakis 2002,
customers
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
34
Chapter 2
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
36
Chapter 2
37
Chapter 2
38
Chapter 2
theory.
Table 2.2. Game models for supply chain coordination
Game type
Model
Key problem
Non-cooperative
Cooperative
Divide
and
align
the
independent
objectives
supply
of
chain
the
additional
total
Pricing game
model
Inventory cost
game model
Pricing and
inventory game
model
Dada
1984;
Choi, 1991;
Minakshi,
1998
Lariviere
and
Srikanth,
1987,
Porteus
2001;
Abad, 2003;
al.,2008
Esmaeili et
2007
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
41
Chapter 2
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
45
Chapter 2
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
47
Chapter 2
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
52
Chapter 3
53
Chapter 3
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
ps
pm
pm
pr
pr
ps
(b) Vertical Nash
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)
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)
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
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)
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)
ps
pm
(a) MR-Stackelberg
ps
s
pm
m
(b) MR-Nash
pm
pr
(c) SM-Stackelberg
pm
s
pr
r
(d) SM-Nash
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)
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
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
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
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)
b 1
3b 3
2b
2b 1
b 2b b 2
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
(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
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
CE I
tends
CE MR S
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
0,
0.
a
a
a
a
a
a
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)
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
75
Chapter 3
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)
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
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)
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
79
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
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
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
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)
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.
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
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
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
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
88
Chapter 4
product l
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
89
Chapter 4
nml : The integer multiplier used to determine the procurement cycle of raw
materials used for product l
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:
X r X r1 X r2
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.
Gml , setup time interval T and procurement decision for raw materials nml
91
Chapter 4
xr xr1 , xr2 ,
j 1,2,..., L
j l
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
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
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
(4.3)
(4.4)
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
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
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
Oml
nml T
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
(4.9)
(4.10)
(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
slv Drl T / Pl
nrl slv Drl T
nlT
Time
Inventory level
(b) Manufacturers inventory level for product l
TDrl / Pl
T / krl
96
Time
Chapter 4
M ax Z m
Gml ,T ,nml
Subject to
Constraints (4.9)-(4.12)
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.
97
Chapter 4
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)
2T 2 hrl Drl
krl * 1 1
Orl
/ 2 .
(4.16)
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
(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)
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)
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.
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
prl
erll
Grl
erll
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
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
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)
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)
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
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:
(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
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
the
manufacturer.
Fixed
x m (0)
find
the
optimal
reaction
104
Chapter 4
achieved. Output the optimal results and stop. Otherwise, x(0) x* , go step 1.
(: 1 , 2 are very small positive numbers)
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
4.6 SUMMARY
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
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
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
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 ,
2 v2
CHAPTER 5
GAME-THEORETIC COORDINATION OF PRICING AND INVENTORY
DECISIONS IN A MULTI-LEVEL SUPPLY CHAIN
5.1 INTRODUCTION
111
Chapter 5
three-level
supply
chain
(supplier-manufacturer-retailer).
Giannoccaro
and
112
Chapter 5
113
Chapter 5
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
Chapter 5
m: Index of manufacturer
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
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
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
xs , xm
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
Supplier
Manufacturer
Retailer
The bottom-level
SS game
The middle-level
MS game
The top-level
RMS game
120
Chapter 5
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
(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
121
Orl krl
T
Chapter 5
Inventory level
TDrl
krl
Order point
T
krl
Time
Inventory level
s TDr
lu
s TDr
lv
Level of product l
TDrl
Pl
T
krl
Time
Time
Ksv T
(c) Supplier vs inventory level of raw material
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)
(5.4)
(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
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
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
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
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
(5.9)
(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
( K sv 1)T slv Drl (Thsv ) + ( K sv 2)T slv Drl (Thsv ) ++ T slv Drl (Thsv ) ,
l
which is equal to
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
hsv
Osv
K sv T
Rsv
(5.12)
Subject to
K sv 1 , 2 , 3, ,. . .
(5.13)
(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:
Subject to
Constraints (5.3)-(5.7)
max Z m
T ,Gml
Subject to
Constraints (5.9)-(5.11)
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.
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
(5.16)
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)
2T 2 hrl Drl
krl 1 1
Orl
/ 2 .
(5.18)
Ar Pl
Grl max 0, l
Gml slv Gsv csv cml
e
v
r
l
Grl
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)
Grl
Cl hrl T
,
2erl 4krl
(5.22)
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
hmrs
v
Sm Om
. (5.23)
T
2 Sm Om
2U m
0 , the optimal T
2
T
T3
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
Ar Pl
Gml max 0, l
Grl slv Gsv csv cml
e
v
r
l
Gml
Arl
(5.26)
(5.27)
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
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
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
K sv
hsv
Osv
K sv T
(5.32)
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
hsv 0 .
(5.35)
Gsv
u 1,...,V
u v
s Gs s El
lu
lv
( K sv 1)T
2 slv erl
2
hsv ,
(5.36)
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
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
5.3.1,
and X sv Gsv , K sv
we
denote
X rl Grl , krl
Chapter 5
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)
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
the
manufacturer.
Fixed
x m (0)
find
the
optimal
reaction
xsv * Gsv * , K sv *
strategy set I sv . If
obtained. Output the optimal results and stop. Otherwise, x(0) x* , go step 1.
132
Chapter 5
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
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
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
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
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
Platform
Product
Manufacturer
Retailer
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
146
Chapter 6
147
Chapter 6
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
(6.2)
(6.3)
(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
Inventory level
TDrl
Pl
T
krl
149
Time
Chapter 6
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
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
l 1
v 1
Ni
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
(6.8)
(6.9)
(6.10)
(6.11)
ij
sv Lij
v 1
Lij Qsv
(6.12)
(6.13)
(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
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
(6.16)
152
Chapter 6
Subject to
pLij 0 ,
Lij Qsv
(6.17)
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:
Subject to:
Constraints (6.2)-(6.6);
M ax Z m ,
xm
Subject to:
Constraints (6.8)-(6.15);
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
153
Chapter 6
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)
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
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
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
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
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
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
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
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
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
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
mn
k n 1
BLmN l Pl
m
max 0, max
,
n Nm
l
m
m
(6.30)
pLmn
BLmnl
,1 n N m 1
min
Nm
l
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
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)
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
x m (0) , using
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
159
Chapter 6
obtained. Output the optimal results and stop. Otherwise, x(0) x* , go step 1.
(: 1 , 2 , 3 are very small positive numbers)
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 ;
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 ,
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
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)
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
(107)
200
80
NCS
WCS
(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)
(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
(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
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
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
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
(107)
6
NCS
WCS
28.0286
5.6451
5.6451
Profit
Profit
28.0286
(107)
26.9132
5.1376
5.05
26.1764
25.3957
19
4.8898
NCS
WCS
100
125
cL12
50
(107)
5.0501
25.2636
24
75
5.0512
25.3957
25.3958
50
5.3383
25
75
100
125
cL12
(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
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
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
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
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
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
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)
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
174
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.
179
Chapter 7
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.
7.3.1 Notations
In this section, we will introduce some other relevant parameters and decisions
variables used for the mathematical models.
Parameters:
181
Chapter 7
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:
ijk : Binary decision variable to indicate whether component Lij has been used to
replace Lik
tsv Lij : Binary decision variable to indicate whether component Lij is supplied by
supplier v
182
Chapter 7
(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
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:
(7.2)
0 Drl Pl
(7.3)
k 1
j 1
ijk
(7.5)
(7.6)
ij
183
Chapter 7
sv Lij
v 1
Lij Qsv
(7.8)
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
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
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
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
i N , v
S ! n S 1!
S N \i
n!
v S
i v S ,
(7.12)
v S
marginal
i v S .
contribution
of
adding
player
to
coalition
is
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 .
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
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
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
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 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
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}
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}
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}
Chapter 7
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
Chapter 7
Chapter 7
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
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
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
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
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 8
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 8
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
Chapter 8
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
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
Chapter 8
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
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
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
0.
b 13
2
2
lim x b lim ln 2
0.
b
b
b
b
3
b
1
b
'
b 1
b 1
2b
b b 3
3b 1
Hence,
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
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
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
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
220
Reference
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
Reference
224
Reference
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
227
Reference
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
230
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