Beruflich Dokumente
Kultur Dokumente
Access to this document was granted through an Emerald subscription provided by 540409 []
For Authors
Downloaded by MAHIDOL UNIVERSITY At 05:24 23 March 2015 (PT)
If you would like to write for this, or any other Emerald publication, then please use our Emerald for Authors service
information about how to choose which publication to write for and submission guidelines are available for all. Please visit
www.emeraldinsight.com/authors for more information.
About Emerald www.emeraldinsight.com
Emerald is a global publisher linking research and practice to the benefit of society. The company manages a portfolio of
more than 290 journals and over 2,350 books and book series volumes, as well as providing an extensive range of online
products and additional customer resources and services.
Emerald is both COUNTER 4 and TRANSFER compliant. The organization is a partner of the Committee on Publication
Ethics (COPE) and also works with Portico and the LOCKSS initiative for digital archive preservation.
T
he spreadsheet environment is superb for Thus the transport buyer needs a tool capable of handling
applications with many variables and com- many variables, complex relationships, and "What if?"
plex relationships. scenarios. The spreadsheet environment is superb for
applications like this. When combined with many of the
commercially available add-in software modules, the
spreadsheet's capabilities encompass linear and non-linear
optimisation, goal seeking, risk analysis, Monte Carlo
simulation, and multiple scenario analysis. Not surprisingly,
Introduction Setting
Knowledgeable transport buyers consider the impact of their The purchase of transport to haul freight involves both
purchases on logistics cost and service elements. This is strategic and tactical transport and inventory decisions.
a challenging task because the interdependence of elements Strategic transport planning encompasses the choice of
such as shipping-time performance, replenishment quantities, mode and the method of acquisition (freight tariffs,
price and freight-rate structures, inventory levels, and contracts, third parties, and lease-buy arrangements),
customer service requirements, form an intricate web that while key tactical planning elements include the selection
complicates the analysis of trade-offs. Such analyses are of specific carriers and the shipment size for specific lanes.
especially complex when uncertain demand and lead time Inventory planning must consider strategic decisions about
conditions prevail. The complexity primarily arises from the the aggregate level and deployment of inventories
task of quantifying the effects of transport performance throughout the logistics network and tactical decisions
attributes on inventory holding costs. Additionally, the about mix of cycle, safety, and pipeline stock for specific
number of variables and calculations needed to produce a links and nodes. Thus an integrated planning approach
solution complicate the analysis. A detailed examination of encompasses both lane and network analyses.
the interactions would be hopelessly impractical without use
of computers and management science models. The microcomputer model presented here applies to the
single-lane scenario, which proceeds as follows. A firm
Received February 1991 stocks a single item to meet uncertain, but stationary and
Revised June 1991, September 1991 independent, customer demand. When the units on hand
TRANSPORT SELECTION: COMPUTER MODELLING IN A SPREADSHEET ENVIRONMENT 29
deplete to a certain level (s), thefirmorders a replacement The transport costs include the direct shipping costs, as
quantity (Q)froma single supplier having multiple carriers well as other kinds of accessorial charges for activities
(from one or more modes) available to transport the order. such as haulage, loading or unloading. The inventory costs
The supplier may represent: (1) a company-owned plant represent the sum of in-transit, cycle and safety stock
or distribution centre, or (2) an independent partner or holding cost components, while the procurement costs
channel ally. Furthermore, each carrier has uncertain represent the sum of ordering (or set-up) costs and the
delivery times and offers an independent schedule of purchase cost. The mathematical terms that calculate the
freight rates. The problem is to determine the transport holding costs of in-transit and safety stock include the
option, the shipment size (Q), and the replenishment level mean (speed) and variance (consistency) of shipping time.
(s) that will minimise the expected total annual costs for This arrangement allows the model to capture the effects
a predetermined level of service. of shipping time performance on holding costs.
transport demand model[4]. The purpose of such "macro macro level inventory theoretic models, the general
level" demand models is either to predict freight flows framework is appropriate for internal analyses. In the
(or market shares) among the modes of transport or to transport planning process, for example, the transport
explain transport choice behaviour among firms[5,6]. Other buyer can easily combine the quantitative analysis of
"micro level" models, however, have adapted this transport alternatives with the appropriate qualitative
framework for the analysis of transport-inventory trade- considerations.
offs within thefirm[7].Although most of these micro level
models concentrate on the single-lane setting, several
models[8,9,10] focus on the logistics network. Assumptions and Refinements
Table I summarises the standard micro level transport
Briefly stated, the inventory theoretic approach defines selection model assumptions and lists the refinements of,
each transport option in terms of speed, on-time and extensions to, these assumptions that are included
consistency, and unit shipping cost. It also defines the in the spreadsheet model. The refinements involve the
product shipped in terms of expected annual transport, characterisation of unit shipping costs and customer
inventory, and procurement cost elements. As Figure 1 service. The conventional approach is to assume that unit
shows, the expected total annual logistics cost for a shipping costs are either constant[12-15] or may be
prespecified level of service and a given shipment modelled adequately by some mathematical function[16].
size is the sum of transport, inventory, procurement costs. One exception is the Stenger et al. [17] model which used
30 IJPD & LM 21,7
a computer program to determine the correct unit shipping Table I. Assumptions, Refinements and Extensions
costs from alternative multi-tier volume freight rate
structures. More recent studies[18-21], however, use a
simple mathematical model that approximates the shipping Standard assumptions
cost per unit by dividing the fixed cost of a fully loaded 1. Single item and lane with multiple transport
vehicle by the units ordered. By contrast, the model options
presented here relies on the power and flexibility of the 2. Stationary, probabilistic and independent
spreadsheet environment to determine unit shipping costs distributions for demand and lead time
precisely and conveniently — even for the traditional multi- 3. Demand during lead time is normally distributed
tier volume freight rate structure that makes unit shipping 4. Continuous review (s,Q) inventory system
costs a discontinuous function of the shipment size. 5. P1 (fraction of cycles experiencing a stockout)
service policy
The second refinement relates to the service control 6. No crossover orders
criterion. The conventional approach sets the desired 7. Freight rates constant or estimated by
probability of no stockout per replenishment cycle (P1 to mathematical model
be the same for each transport alternative. The realised 8. Constant unit cost
P1 will differ from the target, however, because larger
Downloaded by MAHIDOL UNIVERSITY At 05:24 23 March 2015 (PT)
order sizes produce fewer cycles per year and fewer Refinements or extensions
chances of a stockout. Unlike the P1 criterion, the 1. Actual unit shipping costs derived from with
fraction of annual sales not lost (P2), adjusts safety stock multilevel freight rate structures
so that each transport option tested will experience the 2. P2 (fraction of annual sales not lost or
same level of expected annual shortages. Thus the service backordered) service policy
effect is constant for the alternatives evaluated. On the 3. Acquisition costs and multilevel volume discount
other hand, although the order fill (1-P2) criterion is prices
conceptually superior to the cycle stockout (1-1) criterion 4. Shipping speed and consistency as a function of
in transport selection models, the order fill measure entails shipment size
more complexity[22]. Indeed, only Stenger et al.[17] have
tried to incorporate this service element in a micro level
model. As this microcomputer application will illustrate,
however, the spreadsheet environment can readily
encompass the use of the order fill (P2) criterion. assumes that both demand and lead time are uncertain.
Although the dual uncertainty assumption is generally
The spreadsheet model also encompasses two extensions. more realistic, it greatly complicates the task of modelling
The first extension is to add the annual acquisition costs lead time demand accurately[23]. In addition, the
to the total cost equation coupled with the capability to spreadsheet model uses a computer-based approach,
evaluate the impact of multilevel volume discount prices which can easily accommodate the kind of changes
on acquisition costs. The current microlevel models previously discussed that would severely hinder the
assume that the unit cost of an item is constant, which development of elegant analytic models with neat solution
makes the total annual acquisition cost the same for each equations.
transport alternative and thus irrelevant to the solution.
Yet in a channel alliance setting,firmsoften offer quantity
discounts, which can influence the trade-offs when carriers Spreadsheet Design
have independent volume rate structures of their own.
As Figure 2 shows, the display version of the spreadsheet
consists of two major sections. Each section identifies the
The second extension permits the analyst to express the model variables along with their symbols, values, and units
speed and consistency of delivery as a function of the of measure. Cell formulas produce the values shown in
shipment size. For example, the large shipments that by- italics.
pass terminals should experience better shipping time
performance than the small shipments that move through
terminals. Input Section: Model Calibration
The first section of the spreadsheet contains the basic
In addition, the spreadsheet model differs from the best information needed to calibrate the model. This
known microcomputer application[21], ShipSmart, in at information consists of product, inventory, and transport
least two basic ways. The ShipSmart model opts for an elements. The product zone of this section assumes that
analytic approach that rests on two key assumptions: (1) thefirmhas a forecasting system to estimate the standard
that demand is deterministic; (2) that a continuous deviation of period forecast errors (SFE). An estimate
mathematical function can reasonably approximate unit of the standard deviation of daily demand, however,
shipping costs. By contrast, the spreadsheet model would also satisfy this input requirement[24]. In addition,
TRANSPORT SELECTION: COMPUTER MODELLING IN A SPREADSHEET ENVIRONMENT 31
Downloaded by MAHIDOL UNIVERSITY At 05:24 23 March 2015 (PT)
32 IJPD & LM 21,7
B A B C D E F
1 Transport option 0 = LTL truck Worksheet B
2
3 Row/column 0 1 2 3 4
4 Minimum Freight Other Shipping performance
5 weight rate charges Speed On-time
6 0 1 $36.75 $0.00 7.00 1.21
7 1 500 $31.03 $0.00 6.00 1.21
8 2 1,000 $25.85 $0.00 6.00 1.21
9 3 2,000 $22.94 $0.00 5.00 1.21
10 4 5,000 $18.44 $0.00 5.00 1.21
11 5 10,000 $17.58 $0.00 4.00 1.10
12 6 20,000 $12.05 $0.00 4.00 1.05
13 7 30,000 $10.27 $0.00 4.00 1.05
14 8 40,000 $8.98 $0.00 4.00 1.05
Downloaded by MAHIDOL UNIVERSITY At 05:24 23 March 2015 (PT)
15
C A B C D E F
1 Transport option 1 = TL truck Worksheet C
2
3 Row/column 0 1 2 3 4
4 Minimum Freight Other Shipping performance
5 weight rate charges Speed On-time
6 0 30,000 $8.50 $0.25 4.50 1.00
7 1 40,000 $8.50 $0.25 4.00 1.00
8 2 45,000 $8.50 $0.25 4.00 1.00
9 3 45,000 $8.50 $0.25 4.00 1.00
10 4 45,000 $8.50 $0.25 4.00 1.00
11 5 45,000 $8.50 $2.50 4.00 1.00
12 6 45,000 $8.50 $0.25 4.00 1.00
13 7 45,000 $8.50 $0.25 4.00 1.00
14 8 45,000 $8.50 $0.25 4.00 1.00
15
D A B C D E F
1 Transport option 2 = TOFC Worksheet D
2
3 Row/column 0 1 2 3 4
4 Minimum Freight Other Shipping performance
5 weight rate charges Speed On-time
6 0 36,000 $7.75 $0.50 6.00 2.00
7 1 36,000 $7.75 $0.50 6.00 2.00
8 2 40,000 $7.75 $0.50 6.00 2.00
9 3 45,000 $7.75 $0.50 6.00 2.00
10 4 45,000 $7.75 $0.50 6.00 2.00
11 5 45,000 $7.75 $0.50 6.00 2.00
12 6 45,000 $7.75 $0.50 6.00 2.00
13 7 45,000 $7.75 $0.50 6.00 2.00
14 8 45,000 $7.75 $0.50 6.00 2.00
15
this zone contains a unit cost schedule to capture volume- the analyst places the transport data for each option on
sensitive acquisition cost structures. The inventory zone separate, but identically arranged, worksheets (or pages).
offers separate holding cost factors for in-transit and The computer retains these worksheets in memory and
warehouse stock and includes the ordering (set-up) cost saves them as part of worksheet file on a disk. An
and the P2 service criterion (simply identified as P in "©Index" function uses the transport selection code in
Figure 2).
cell G6 to pull up the correct information. This
The transport cost-performance table takes advantage arrangement not only organises data entry, but makes
of the three-dimensional capability provided by the latest it possible to automate the multiple "What if?" analyses
versions of spreadsheet software. As Figure 3 illustrates that form a key part of the model's output section.
TRANSPORT SELECTION: COMPUTER MODELLING IN A SPREADSHEET ENVIRONMENT 33
Output Section: Analysis and Solutions input section of the worksheet. As previously indicated,
The second section of the spreadsheet represents the heart this information is important because transit-time
of the model. The "Elements" and "Expected annual cost" performance affects the in-transit and safety stock
zones contain the working equations, which provide a calculations.
detailed study of the impact of a transport option on total
logistics costs for preset values of Q and P. The "Solution Inventory equations. The most complex task involves the
table" enumerates the expected total annual logistics cost computation of safety stock. The procedure entails
(TAC) of each transport option over a range of Q values. calculations for:
The specific equations producing the italicised values in this (1) the standard deviation of lead time forecast errors
section are shown in Figure 4. Since the foundation for these (SDLTFE);
equations is well established in the logistics production
management literature, a brief survey of the key aspects (2) the partial expectation (E(Z)), which represents the
of the transport, inventory and procurement working and expected units short per replenishment cycle per
cost equations should suffice[22,25]. unit of SDLTFE;
(3) the directly corresponding safety stock multiplier (Z).
Transport equations. The computer model examines the The model makes these calculations as follows. First, the
"weight breaks" to identify the correct freight rate and standard compounding technique is used to determine
Downloaded by MAHIDOL UNIVERSITY At 05:24 23 March 2015 (PT)
minimum weight for any order size Q. This step enables SDLTFE from the parameter estimates of demand (DD
the model to determine the correct "billed" weight as either and SFE2) and transit time (LT and VARLT)
the actual weight or the minimum weight and then calculate distributions[26]. Second, the partial expectation equation,
the true shipping cost per order. The model uses the same which assumes the complete lost sales case, is easily
method to identify the speed and on-time performance determined from other input values[22]. Third, the model
values in the "Transport cost-performance table" in the uses Brown's[27] rational approximation method to
Model Equations
F
A B C G H
Results given Q and P
Elements Expected annual costs
Transport Transport Q
28 Shipment weight SW (Q*UW) Direct shipping SC +BW*0.01* RATE *R/Q
29 Minimum weight MW @VL00KUP(SW,TABLE,1) Other charges AC (R*UW*0.01*CHGS)
30 Billed weight BW @MAX(SW,MW) Total TRN @SUM(H28..H29)
31 Freight rate RATE @VLOOKUP(SW,TABLE,2)
32 Other charges CHGS @VLOOKUP(SW.TABLE,3)
33 Speed (mean periods) LT @VLOOKUP(SW.TABLE,4)
34 On-time (variance) VARLT @VLOOKUP(SW,TABLE,5)
35
36 Inventory Inventory
37 Units/period DD (R/PY) Cycle stock CSC + CS*(US+TRN/R)*IW
38 Unit cost UC @VLOOKUP(Q,PRICES,1) In-transit stock ISC + IS*(US+TRN/R)*IT
39 Cycle stock CS (Q/2) Safety stock ssc + SS*(UC+TRN/R)*IW
40 In-transit stock IS @VL00KUP(SW,TABLE,4)*DD Total INV @SUM(H37..H39)
41 Safety stock SS (Z*SDLTFE)
42 Std.dev.LTD SDLTFE @SQRT(LT*SFE2 + DD2*VARLT)
43 Partial expectation E(Z) ((1-P)/P)*(Q/SDLTFE>
44 Multiplier Z @IF(C43<0.398,(C43-0.39894228)
*(-1.75294+0.4442135 *C43
-(0.07061455*C432)
-0.17592241/C43+0.044212641)
-0.0012267386/(C43 + 0.00030570313)),0)
Procurement Procurement
47 Reorder level s (LT*DD)+SS Ordering cost OC (R/Q*A)
48 Reorder quantity Q 444 Purchase PUR +R*@VL00KUP(Q,PRICES,1)
49 Cycles per year RY +R/Q Total PRO @SUM(H47..H48)
50 Total costs TAC @SUM(TRN,INV,PR0)
Table references:
Table - F11..K19 keys on weight breaks to lookup transport cost and performance elements
Prices - B11..C14 keys on order size to lookup unit price
34 IJPD & LM 21,7
References
1. Starfield, A.M., Smith, K.A. and Bleloch, A.L., How to
Model It: Problem Solving for the Computer Age, McGraw-
Hill, New York, NY, 1990.
2. Lilien, G.L., Marketing Mix Analysis with Lotus 1-2-3,
The Scientific Press, Palo Alto, California, 1987.
3. Troutt, M.D., Tadisina, S.K. and Clinton, R.J.,
"Interactive Optimization Aspects of Electronic
Spreadsheet Models for Design and Planning'', Journal
of the Operational Research Society, Vol. 42 No. 5, 1991,
pp. 349-56.
4. Baumol, W.J. and Vinod, H.D., "'An Inventory Theoretic
Model of Freight Transport Demand", Management
Science, Vol. 16 No. 7, 1970, pp. 413-21.
Downloaded by MAHIDOL UNIVERSITY At 05:24 23 March 2015 (PT)
15. Larson, P.D., "The Economic Transportation Quantity", 22. Silver, E.A. and Peterson, R., Decision Systems for
Tranportation Journal, Vol. 28 No. 2, 1988, pp. 43-8. Inventory Management and Production Planning, Wiley,
New. York, NY, 1985, p. 284.
16. Langley, C.J. Jr, "The Inclusion of Transportation Costs
in Inventory Models: Some Considerations", Journal of 23. Bagchi, U., Hayya, J.C. and Ord, J.K., "Modelling
Business Logistics, Vol. 2 No. 1, 1981, pp. 106-25. Demand during Lead Time", Decision Sciences, Vol. 15,
1984, pp. 157-76.
17. Stenger, A.J., Coyle, J.J. and Prince, M.S.,''Incorporating
24. Zinn, W. and Marmorstein, H., "Comparing Two
Transportation Costs and Services into the Inventory
Alternative Methods of Determining Safety Stock Levels:
Replenishment Decision", in House, R.G. and Robeson,
The Demand and the Forecast Systems", Journal of
J.F (Eds), Applied Distribution Research, Ohio State
Business Logistics, Vol. 11 No. 1, 1990, pp. 95-110.
University, Columbus, Ohio, 1977, pp. 22-6.
25. Bowersox, D.J., Closs, D.J. and Helferich, O.K., Logistical
18. Hall, R.,' 'Dependence between Shipment Size and Mode Management, 3rd edition, Macmillan, New York, NY, pp.
in Freight Transportation", General Motors Research 211-4.
Laboratories Report GMR-4438, Warren, MI, 28 July
1983. 26. Mentzer, J.T. and Krishnan, R., "The Effect of Normality
on Inventory Control/Customer Service", Journal of
19. Blumenfeld, D.E., Hall, R.W. and Jordan, W.C., "Trade- Business Logistics, Vol. 6 No. 1, 1985, pp. 101-20.
Downloaded by MAHIDOL UNIVERSITY At 05:24 23 March 2015 (PT)
John E. Tyworth is Associate Professor of Business Logistics at Pennsylvania State University, University Park, PA, USA.
This article has been cited by:
1. Abraham Mendoza, José A. Ventura. 2013. Modeling actual transportation costs in supplier selection and order quantity
allocation decisions. Operational Research 13, 5-25. [CrossRef]
2. Abraham Mendoza, José A. Ventura. 2009. Estimating freight rates in inventory replenishment and supplier selection decisions.
Logistics Research 1, 185-196. [CrossRef]
3. Gust Blauwens, Nico Vandaele, Eddy Van de Voorde, Bert Vernimmen, Frank Witlox. 2006. Towards a Modal Shift in Freight
Transport? A Business Logistics Analysis of Some Policy Measures. Transport Reviews 26, 239-251. [CrossRef]
4. Scott R. Swenseth, Michael R. Godfrey. 2002. Incorporating transportation costs into inventory replenishment decisions.
International Journal of Production Economics 77, 113-130. [CrossRef]
Downloaded by MAHIDOL UNIVERSITY At 05:24 23 March 2015 (PT)