You are on page 1of 7

ARTICLE IN PRESS

Robotics and Computer-Integrated Manufacturing 22 (2006) 550556


www.elsevier.com/locate/rcim

Economic value added of supply chain demand planning:


A system dynamics simulation
J. Ashayeria,, L. Lemmesb
a
Center for Economic Research (CentER), Department of Econometrics and Operations Research, Tilburg University, The Netherlands
b
EyeOn ConsultancyEindhoven, The Netherlands
Received 7 October 2005; accepted 14 November 2005

Abstract

The development of a successful demand plan is typically a joint effort between different functional units such as Logistics, Marketing,
Sales and executive management at one hand and between different business units on the other. Starting a project to structurally improve
the demand planning often requires convincing all parties involved in such an effort. The key is to quantify the bottom-line impact of an
increased demand planning reliability in the supply chain. This paper proposes a system dynamics simulation modeling framework that
allows different managers to examine how improvements in their demand reliability will impact the overall corporate bottom-line. For
example, supply chain managers can investigate how proposed changes in the supply chain demand forecasting structure, different
suppliers, different logistics routes, or alternative inventory methods, may increase the overall protability. The simulation model has
been tested, validated with a real-life case of LG. Philips Displays Europe.
r 2006 Elsevier Ltd. All rights reserved.

Keywords: Supply chain; Demand reliability; System dynamics; Simulation

1. Introduction related consulting services, demand planning in complex


supply chains is remained a constant challenge due to
Demand planning is a critical business process that market dynamics. In order to exploit the markets and
impacts all aspects of supply chain management, and the improve the bottom-line, historically acceptable forecast
business bottom-line [1]. The ever-increasing trend in accuracy level (around 50%) is not tolerable. The rough
globalization of business has also made the demand monthly forecasting cannot provide adequate support for
planning and management more complex due to the fact demand management. And since demand planning in
that the markets are becoming progressively more un- scattered throughout the chains, it becomes more compli-
predictable, fragmented, and dynamic. These changes cated. Companies in different industries are facing this
requires for a supply chain managers to manage dynami- problem, however the electronic industry is hit most.
cally and choose best demands, instead of simply generat- A demand forecast will be never perfect. It is certain that
ing some rough demand projections based on static always there are forecast errors present. Since marketing
analysis of aggregated market data. For many products, and sales normally will have the most information on
especially in electronic sector the combination of decreas- market trends and events and on major accounts, it is
ing product life cycles and intolerant markets has increased extremely important that a leading change agent in the
the cost of error of choosing to supply the wrong demand. chain tries to improve sales-planning reliability conjunctly
Increasingly, companies are implementing the best-of- with them. But how they would think about demand
the-bread ERP and advanced planning systems (APS). planning in a structured manner? After all, marketing and
Despite millions of dollars invested in ERP and APS, and sales entities give priority to increasing or consolidating
market share and not so much to the actual planning.
Corresponding author. Tel.: +31 13 466 2176; fax: +31 13 466 3280. Through proper demand planning, the forecast error can
E-mail address: j.ashayeri@uvt.nl (J. Ashayeri). become smaller; and by early predicting the market

0736-5845/$ - see front matter r 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.rcim.2005.11.010
ARTICLE IN PRESS
J. Ashayeri, L. Lemmes / Robotics and Computer-Integrated Manufacturing 22 (2006) 550556 551

dynamics, a more reliable sales plan can be generated. The chain using SD simulation and discuss issues that have
inuence of a more reliable sales plan is not negligible. evolved, and presents a taxonomy of research and
Amongst many benets the following are the highlights: development in SD modeling in supply chain management.
Of the early work we can refer to Towill et al. [6] and
1. The sales volume increases, because of the higher Wikner et al. [7] who simulated different supply chain
customer service level. improvement strategies on demand amplication. Ashayeri
2. Reduction of the operational costs, like less obsolete and Keij [8] model a large European distribution chain
inventory and less capital costs for holding inventory, using SD simulation. Affeldt [9] describes why simulation
less obsolete inventory and a lower required safety and why SD. He argues that SD simulation illuminates for
inventory. its users: systemic feedback loops, systemic delays, and
3. The weighted average cost of capital (WACC) percen- unintended consequences. He adds that development of a
tage decreases. The company will have a lower risk simple, yet relatively robust, causal loop diagram (CLD)
prole for shareholders leading to a lower WACC encapsulates a portion of the business in which these three
percentage. effects were uncontrolled, and therefore shows the real
business (market) dynamics.
The goal of this paper is to construct a generic A general discussion on supply chain using simulation
simulation model that measures the bottom line nancial can be found in Bansal [10]. They identify the potential
consequences of an improved demand planning reliability. promise of simulation domain and provide a brief review of
The term generic refers to the fact that different supply this domain and modeling methodologies as applied to
chains can use this model to measure the nancial supply chain systems. Sterman [11] details the SD applica-
consequences as an effect of improved demand (sales) tions in dynamic business environments. Ganeshan [12]
planning reliability for their supply chain. It is crucial to studies the impact of selected inventory parameters on the
point out the added value of such a tool is in processing the performance of an expanded and comprehensive retail
results in a language understood by a typical marketing supply chain using simulation. The study concludes that
and sales representative. This way, in some cases, market- information sharing between echelons in the supply chain
ing and sales can even function as a trigger for the changes yields a higher level of service. Ritchie-Dunham et al. [13]
needed in the supply chain. describe a simulation game designed to quantify the
The remainder of the paper is organized as follows. benets of an ERP system coupled with the balanced
Section 2 reviews briey the relevant literature on supply scorecard framework in an extended enterprise. Results
chain simulation using system dynamics (SD) approach. from this game outlines the benets of strategies such as
Section 3 explains few details of the simulation model. systems integration, data and process standardization,
Section 4 describes the results expected from the simulation visibility across the business enterprise, improved decision
using data from a real-life case situation, and nally support system. Anderson et al. [14] present a similar paper
Section 5 presents the conclusions. for service supply chain environment.
More recent works include Pundoor and Herrman [15],
2. Literature review who present a simulation framework that follows the
Supply Chain Operations Reference (SCOR) Model. The
The use of SD modeling in supply chain has been very simulation approach is rather limited as it is discrete and
limited but recently given complexity in supply chains has spreadsheet based. However, it is useful for management
gained increased popularity. The dynamic nature of supply training and development. Umeda and Lee [16] describe a
chain systems and their behavior depends on the un- design specication for a generic, supply-chain-simulation
certainties of customers demand, different suppliers, system. The proposed simulation system is based on
different logistics routes, or alternative inventory methods, schedule-driven (pull) and stock-driven (push) control
etc. In fact uncertainty rules the supply chain. Therefore, it methods to support the supply chain management. The
is natural to apply SD simulation. Perhaps the most well approach is also discrete-event simulation and does take
known supply chain being simulated using SD concepts is into account the hidden dynamics of supply chain. Kamath
the Beer Game, dating back to early 1960s. Of the more and Roy [17] present a system SD based experimental
contemporary works in the literature one can refer to the method for designing a supply chain structure for a volatile
Cooperative Supply-Chain System Dynamic Process Flow market of short lifecycle product. Here the capacity
Model (DPFM) [2]. constrained supply chain is simulated to improve product
Over the last decade, more attention is paid to the use availability and customers service.
SD simulation for analyzing different aspects of supply Other related studies include Swaminathan et al. [18]
chain management. Given the limited space available, here who discuss a supply chain simulation modeling using the
only very few papers are referenced. For a more recent SD concepts and agent-based simulation framework.
overview of literature see Terzi and Cavalieri [3], Kleijnen A similar approach is the work of Schieritz [19], Schieritz
[4], and Angerhofer and Angelides [5]. The latter paper and GroXler [20], and Borshchev and Filippov [21].
gives an overview of research work in modeling supply Baumgaertel and John [22] develop also an agent-based
ARTICLE IN PRESS
552 J. Ashayeri, L. Lemmes / Robotics and Computer-Integrated Manufacturing 22 (2006) 550556

approach for highly efcient simulation of supply networks


in which several nodes use advanced planning and
Suppliers Factories Regional Sales Customers
scheduling (APS) systems. distribution Organizations
Centers
3. The simulation model

This section describes the proposed supply chain


simulation system. The model has been set up to simulate
a supply chain from supplier to customer, and to give
an insight in the nancial aspects of this supply chain (see Raw Finished
Fig. 1). The supply chain, the statistical input parameters material product
and the simulation parameters are input requirements for
the model, which implemented in Excel. By running the
model, which is implemented in Powersim, for different Fig. 2. General set-up of the supply chain.
business scenarios, the scenarios can be analyzed and the
simulation parameters adjusted. The eventual results are 5. Transportation, which models transportation time delays
worked out in graphs and tables in Excel and form the and costs.
basis of management decisions. 6. Inventory planning, models inventories, computes re-
commended inventory levels, safety stocks at various
3.1. Set-up supply chain locations.
7. Suppliers, which supply raw materials.
The model can be used for diverse businesses with 8. Supply chain structure, which provides the allocation of
completely different supply chains. This supply chain production and distribution resources to forecast
normally is to be set-up once in the model, including lead demand under capacity and supply constraints.
times, order intervals, safety stocks and other logistical
parameters. The model also offers the possibility to Each of functions in the simulation model is character-
simulate with the supply chain by, for example, adding, ized by a set of static parameters and a set of what-if
moving or closing a factory or a sales organization. The (dynamic) parameters that affect a modeled supply chain.
model developed provides eight types of function/player in Given the supply chain structure and the parameters, the
a supply chain (see Fig. 2). nancial consequences are calculated.
The supply chain can be customized to reect different
1. Customers, which provides customer orders. businesses. The number of factories, distribution centers
2. Sales organizations (marketing/sales), which models and sales organizations can be modeled as well as lead
product forecasts including uctuating demand. times, order intervals, safety stocks and other logistical
3. Factories, which models assembly delays and tracks raw parameters. Static parameters like product characteristics
material (input) and nished goods (output). and some xed cost parameters are set prior to simulation
4. Distribution, which models (regional) distribution cen- and these are xed for the different simulation runs. The
ters, including nished goods inventory. what-if parameters can be changed between different
simulation runs in order to clarify the nancial conse-
quences of certain actions. Although it is possible to
Define Define change the supply chain set-up or different static para-
Set up
Excel

supply chain
statistic input simulation meters, the demand planning reliability has been the focus
parameters parameters of this study. Given the demand planning settings, different
scenarios can easily be simulated for instance:

Redefine  An improvement of demand planning reliability for the


Environment

Model simulation
Powersim

entire organization with x percent.


parameters
 Demand planning reliability improvement of the worst
performing sales organization up to the organization
Scenario average.
analysis  An improvement in demand planning reliability for only
certain category of products with y% (e.g. fast movers
A class products).
Excel

Results
The model calculates the bottom-line nancial conse-
quences of the different scenarios. Results are reported in
Fig. 1. General structure of the simulation model. terms of the economic value added (EVA), which is
ARTICLE IN PRESS
J. Ashayeri, L. Lemmes / Robotics and Computer-Integrated Manufacturing 22 (2006) 550556 553

determined by subtracting the capital cost from the net


operating prot after taxes. This is a very recognizable format Realised demand vs Planneddemand
for the upper management that in general is nancially
oriented. Expected costs for improvement projects can be
compared with the yearly nancial gains. This way a rational xi xi
trade-off can be made on the start of planning improvement REL = 1
i
*100%
projects and improvement priorities can be set. xi
i
When the nancial consequences of a specic scenario
are expected to be higher than the anticipated project costs
of realizing this scenario, then there is a healthy business
case for starting up the project. Note that the project Demand Planning Reliability
costs will be a one off expenditure whereas the nancial
consequences will reiterate on a yearly basis. Fig. 3. The denition of sales planning.
As illustrated in Fig. 1, useful results will only be
achieved after going through a number of scenario analyses,
including the redenition of simulation parameters. In the sales planning reliability the realized sales are calculated,
simulation model a supply chain can be tailor made to the taking into account the stock on hand. This information is
specics of a company. We consider (see Fig. 2) a ve-tiered used for generating internal orders of nished products and
supply chain network consisting of: suppliers, factories, for external orders of raw materials. This way, stock levels
regional distribution centers, sales organizations, and custo- and part of the capital costs are calculated, updated and
mers. These components are dened by several attributes simulation advances according to the lead-time functions.
such as lead-time, lead-time variability, nancial indicators, The nancial consequences of each simulation step are
logical and dynamic relationships, and other logistics allotted in the structure of the EVA tree (see Fig. 4). This
parameters within the simulation model. structure is recognizable for top-management and com-
bines the prot and loss account and the balance sheet.
3.2. Define statistical input parameters
4. Results: A case study
In the model a number of parameters will be dened that
are basically independent of the different business scenar- In this section the results expected from the simulation
ios. For example product characteristics: product location model are outlined using the data from a real-life case. Due
and when applicable the Pareto distribution of products to space limitation, extensive elaboration of all simulation
(8020 rule), the sales plan and the expected selling and scenarios and output is impossible. Therefore, we briey
purchasing prices for the next 24 months. A number of discuss the main output.
xed and variable cost elements and a number of nancial LG. Philips Displays is a large supplier of tubes for use
parameters, among others the expected payment periods of in televisions and computer monitors. The company was
incoming and outgoing invoices and the average cost of formed in 2001 and is the joint venture between the
capital (WACC %) are included in the model. The model Cathode Ray Tube (CRT) businesses of LG Electronics of
also offers the opportunity to simulate with parameters by, Korea and Royal Philips Electronics of the Netherlands.
for example, taking another Pareto division or changing LG. Philips supplies their tubes to amongst others global
the production location for a product. set- and monitoring making companies, like Philips TV,
Loewe and Bang & Olufsen. LG. Philips Displays produces
3.3. Define simulation parameters two different tubes, color picture tubes (CPTs) and color
display tubes (CDTs). The company has a big market share
The key parameter used for simulation in the different of the CPT supplies in Europe; one-third of the European
scenarios is the demand (sales) planning reliability. This television and more than half of the European color display
parameter is calculated at SKU level, in order to take the tube markets is supplied by LG. Philips Displays. This
unreliability in the product mix into account. The denition investigation is focused only on the CPTs produced and
of sales planning reliability as used in the model is 1 minus the sold in Europe. The products that are imported in Europe
sum of the absolute deviations of realized sales and planned are not taken into this study.
sales divided by the sum of planned sales. This calculation is In this study, we considered all six CPT-manufacturing
made at SKU level, because in the model we need to make a facilities in different European countries along with
distinction of the consequences for stock levels (see Fig. 3). regional distribution centers, sales organization units and
all customers. Several scenarios were studied and the
3.4. The model results were analyzed to nd out the best of improving
demand (sales) reliability in the chain.
In the model the above-mentioned parameters are used The results of the different demand (sales) planning
as input. Based on the expected sales plan and the expected reliability scenarios are presented in the EVA structure.
ARTICLE IN PRESS
554 J. Ashayeri, L. Lemmes / Robotics and Computer-Integrated Manufacturing 22 (2006) 550556

Sales

Contribution
Margin
Cost of
Goods sold EBIT

Cost of Net Operating


Profit After
Organisation Taxes
Taxes
EVA

WACC

Capital Cost

Woring
Capital

Fig. 4. Economic value added (EVA) tree.

Table 1 Economic Value Added vs. Demand planning reliability


EVA improvement in Euro (Mln)
The output in excel 5

Total per period 4


Revenues
Cost BOM 3
Prot not realized
Contribution margin
Fixed cost 2
Variable cost total
Variable transport cost 1
EBITA (earnings before interest)
Taxes 0
NOPAT (net operating prot after tax) 70 71 72 73 74 75 76 77 78 79 80
Capital cost Demand planning percentage
EVA total 70% =Base scenario
Different costs Fig. 5. Example 1 standard results of the simulation model.
Revenue loss total
Out-of-stock revenue loss
Incourant revenue loss
Phasing-out revenue loss project costs will be a one off expenditure whereas the
Variable cost total nancial consequences will reiterate on a yearly basis.
Variable personnel cost The bottom-line results is generated based on the
Other variable cost
Damage-and-theft theft cost
output of a simulation run and is made in Excel. Every
Fixed cost total simulation the output is calculated. This output is
Fixed personnel cost displayed in Table 1.
Other xed cost The rst simulation is made for the current situation for
Capital cost total the validation purpose. This is the situation with the actual
Capital cost capital investment
Inventory cost
sales planning reliability of the company. This is the
Accounts payables reference run. The simulation can then be done for
Accounts receivables different scenarios and can be compared to the reference
run.
As illustrated in Fig. 5, useful results will only be
achieved after going through a number of scenario
This way we can make a clear link between supply chain analyses, including the redenition of simulation para-
practice and the more nancially oriented top manage- meters. This graph shows the EVA improvement in
ment. When nancial consequences of a specic scenario millions of Euros in relation to the improvement of the
are higher than the anticipated project costs of realizing the sales planning reliability. The star in this example shows
scenario, there is a healthy business case for starting up that with an improvement of sales planning reliability from
the project. What has to be taken into account is that the 70% to 73% the EVA will improve by h2.5 million yearly.
ARTICLE IN PRESS
J. Ashayeri, L. Lemmes / Robotics and Computer-Integrated Manufacturing 22 (2006) 550556 555

C 4.4M Delta met Basis scenario

Sales C 2.7M

Contribution
C 1.7M Margin C 2.4M
Cost of
Goods sold C 0.3M EBIT C 1.8M

Cost of Net Operating


Organisation C 0.6M Profit AfterTaxes
Taxes + C 2.5M

EVA

10%

WACC
C 0.7M

Capital Cost
+ C 7M
Working
Capital

Fig. 6. Example 2 standard results of the simulation model.

Burns and Ulgen [23] in a recent paper on delays in EVA improvement in relation to the base scenario
dynamic supply chain systems use causal loops and Demand planning reliability +3% point
3
substantiate that reduction of delays in the transmission
of information and in the delivery of goods along supply
chains, coupled with changes in inventory ordering policy
EVA improvement in Euro (Mln)

can substantially reduce the amount of inventory in the Less capital costs
pipeline and save rms millions of dollars. Here we can
2
clearly quantify such issues and present results that has Less stock costs
been rarely seen at the top management level. By
quantifying costs and returns in very detail, we can Less price erosion
measure how, for example, demand-planning reliability
improvement can increase EVA (see Fig. 6). In this EVA Less depreciation
1 obsolete stocks
tree per element the positive delta is calculated in relation
to the basic scenario. In this example, the EVA improve- Less loss of sales
ment of h2.5 million of Fig. 5 is further illustrated. This
would allow the decision maker to choose the right
option(s) for improving the supply chain performance.
0
Another way to present the results is shown in Fig. 7. In EVA improvement
this chart per underlying business issue the positive deltas
in relation to the base scenario are given. Here the EVA Fig. 7. Example 3 standard results of the simulation model.
improvement of h2.5 million of Fig. 5 is further analyzed.
Overall the simulation results show that by different
demand (sales)-planning improvement strategies the cur-
rent planning reliabilities (ranging from 28% to 87% for of the operational costs is a consequence of a reduction of
different sales organization units, and factories) can be the inventory holding costs. The WACC percentage is not
increased to 60% with a maximum limit of 90%, resulting simulated in the model, so some assumptions with respect
in the EVA to increase, approximately with h4.6 million. to this percentage are made. In the extreme case, where the
The improvement of the sales-planning reliability has an sales planning reliabilities for the different manufacturing
inuence on at least three important facts, the delivery plants are improved with 60% and the WACC decreased
service provided to the customer, reduction of the from 11.7% to 10%, the EVA increases with almost h6
operational costs and the decrease of the WACC percen- million with respect to the current situation.
tage. Note these results are derived from a large number of
The higher customer service level results in a decrease of simulation runs after reaching the steady-state situation.
stock-outs and consequently less lost sales. The reduction Although most of these numbers are not the exact savings
ARTICLE IN PRESS
556 J. Ashayeri, L. Lemmes / Robotics and Computer-Integrated Manufacturing 22 (2006) 550556

rather an estimation of them, these estimates show the [5] Angerhofer BJ, Angelides MC. System dynamics modeling in supply
value of the proposed simulation model as a powerful tool. chain management: research review. In: Joines JA, Barton RR, Kang
K, Fishwick PA, editors. Proceedings of the 2000 winter simulation
conference, 2000.
5. Conclusions [6] Towill DR, Naim MM, Wikner J. Industrial dynamics simulation
models in the design of supply chains. Int J Phys Distrib Logist
Demand planning has emerged as an increasingly Manage 1992;22(5):313.
[7] Wikner J, Towill DR, Naim M. Smoothing supply chain dynamics.
important issue for supply chain. The challenge of demand
Int J Prod Econ 1991;22(3):23148.
management is to identify market dynamics in advance. In [8] Ashayeri J, Keij R. Global business process re-engineering: a system
supply chains, higher accuracy performance measures dynamics based approach. Int J Oper Prod Manage 1998;18:81731.
require more rigorous planning of demand. Demand is [9] Affeldt JF. The application of system dynamics (SD) simulation to
the primary driver in both macro supply network design enterprise management. In: Farrington PA, Nembhard HB, Sturrock
DT, Evans GW, editors. Proceedings of the 1999 winter simulation
and operations. In network design, optimization tools are
conference, 1999.
used to perform extensive sensitivity analysis to ensure that [10] Bansal S. Promise and problems of simulation technology in SCM
over-investment in capacity and facilities does not occur. domain. In: Yucesan E, Chen C-H, Snowdon JL, Charnes JM,
This paper examined dynamic simulation for ner analysis editors. Proceedings of the 2002 winter simulation conference, 2002.
of the market dynamics impact on the bottom line. [11] Sterman JD. Business dynamics, systems thinking and modeling for a
complex world. McGraw-Hill Higher Education; 2000.
In these times of economic downturn every Euro [12] Ganeshan R. The impact of inventory and ow planning parameters
invested will have to be accounted for beforehand (which on supply chain performance: an exploratory study. Int J Prod Econ
obviously is a good business attitude in better times too). 2001;71(1-3):11122.
This also holds for demand planning improvement [13] Ritchie-Dunham J, Morrice DJ, Anderson EG, Scott J. A strategic
projects. It would be wrong though to postpone invest- supply chain simulation model. In: Joines JA, Barton RR, Kang K,
Fishwick PA, editors. Proceedings of the 2000 winter simulation
ments that repay themselves quickly. Worse even is to not
conference, 2000.
invest in them at all out of extreme carefulness. Higher [14] Anderson EG, Morrice DJ. A Simulation game for service-oriented
operational turnover, higher margin, and lower operational supply chain management: does information sharing help managers
costs are, after all, nancial results that the company will with service capacity decisions? J Prod Oper Manage 2000;9(1):
see in the operational results every year, on a continuous 34251.
[15] Pundoor G, Herrmann JW. A hierarchical approach to supply chain
basis. On the other hand a complex project with minimal
simulation modeling using the Supply Chain Operations Reference
nancial results is to be of a certain strategic magnitude in model. Department of decision and information technologies. Robert
order to be initiated. In such a case it might be wise to H. Smith School of Business, University of Maryland, College Park,
focus management attention on something else. MD 20742, USA, 2004.
With the simulation model discussed here the nancial [16] Umeda S, Lee YT. Design specications of a generic supply chain
simulator. In: Ingalls RG, Rossetti MD, Smith JS, Peters BA, editors.
consequences of an improved demand (sales) planning
Proceedings of the 2004 winter simulation conference, 2004.
reliability can be calculated in a graspable way. The model [17] Kamath NB, Roy R. Supply chain structure design for a short
provides the opportunity to simulate the consequences of lifecycle product: a loop dominance based analysis. Proceedings of
different scenarios such as changes I inventory policy, the 38th Hawaii international conference on system sciences, 2005.
customer service level, or allocation structure of supply [18] Swaminathan JM, Smith SF, Sadeh NM. Modeling supply chain
dynamics: a multiagent approach. Decision Sci 1998;29(3):632707.
chain. This way the right priorities can be determined for
[19] Schieritz N. Integrating system dynamics and agent-based modeling,
possible improvement projects. The model application Mannheim University Industrieseminar, Schloss D-68131 Mannheim,
proved the value of SD simulation in a real-life setting as Germany, 2002.
indispensable tool. [20] Schieritz N, GroXler A. Emergent structures in supply chainsA
study integrating agent-based and system dynamics modeling. In:
Proceedings of the 36th Hawaii international conference on system
References sciences, 2003.
[21] Borshchev A, Filippov A. From system dynamics and discrete event
[1] Lee HL. Ultimate enterprise value creation using demand-based to practical agent based modeling: reasons, techniques, tools. XJ
management. Stanford Global Supply Chain, Management Forum Technologies www.xjtek.com, Russia, 2004.
September, SGSCMF-W1-2001, 2001. [22] Baumgaertel H, John U. Combining agent based supply net
[2] Hillier FS, Lieberman GL. Introduction to operations research. New simulation and constraint technology for highly efcient simulation
York: McGraw-Hill Publishing Company; 1990. of supply networks using APS systems. In: Chick S, Sanchez PJ,
[3] Terzi S, Cavalieri S. Simulation in the supply chain context: a survey. Ferrin D, Morrice DJ, editors. Proceedings of the 2003 winter
Comput Ind 2004;53(1):316. simulation conference, 2003.
[4] Kleijnen J. Supply chain simulation tools and techniques: a survey, [23] Burns JR, Ulgen O. Removing information delays from supply
Center for Economic Research (CentER). The Netherlands: Tilburg chains: the implications, Internet reference: http://www.pmcorp.com/
University; 2004. pub_simulation.shtmwww.pmcorp.com/pub_simulation.shtm, 2004.

You might also like