See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/236455073

Supply chain simulation methods analysis: an

application to the Beer Game

Conference Paper · October 2010

CITATIONS READS

0 650

4 authors:

Daniel Guimarans Julija Petuhova

Amsterdam University of Applied Sciences/C… Riga Technical University
52 PUBLICATIONS 138 CITATIONS 16 PUBLICATIONS 44 CITATIONS

SEE PROFILE SEE PROFILE

Yuri Merkuryev Juan José Ramos

Riga Technical University Autonomous University of Barcelona
129 PUBLICATIONS 493 CITATIONS 33 PUBLICATIONS 64 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Air transport delay propagation and disruption mitigation View project

Simulation and Analysis of Container Freight Train Operations at Port Botany View project

All content following this page was uploaded by Daniel Guimarans on 12 August 2014.

The user has requested enhancement of the downloaded file.

SUPPLY CHAIN SIMULATION METHODS ANALYSIS:
AN APPLICATION TO THE BEER GAME
Daniel Guimarans(a), Julija Petuhova(b), Yuri Merkuryev(c), Juan José Ramos(d)

(a), (d)
Telecommunication and System Engineering Dept., Universitat Autònoma de Barcelona, Spain
(b), (c)
Department of Modelling and Simulation, Riga Technical University, Latvia
(a)
[email protected], (b)[email protected], (c)[email protected], (d)[email protected]

ABSTRACT (b) Discrete-Event Systems Simulation: event

The Beer Game has a typical supply chain structure that or process oriented simulation, etc.
permits exploring a variety of supply management
concepts. Many modelling methods have been used for Analytical models have their place at a tactical
supply chain analysis and so they can be applied to the level in the design of supply chains. Analytical
Beer Game specific case study. Among them, discrete- techniques are able to solve batch sizing and job
event systems simulation has deserved special attention sequencing problems, yet fail to throw much light on
due to its suitability for modelling dynamic systems the dynamic behaviour of the supply chain as a whole
with a high degree of detailed elaboration and stochastic (Riddals, Bennett and Tipi 2000). Analytical techniques
factors. For this reason, several discrete-event are useful in providing solutions to local tactical
simulation oriented models have been elaborated to problems. Nevertheless, the impact of these solutions on
tackle the Beer Game and, by extension, multi-echelon the global behaviour of the whole supply chain can only
supply chains. In the present paper, four of these models be assessed using dynamic simulation. In addition, the
are described. Some of their applicability characteristics computational burden associated to such techniques is
are also outlined, so a further discussion of their to be considered an important drawback.
suitability according to simulation purposes can be Ishii, Takahashi and Muramatsu (1988) developed
done. Conclusions extracted from this analysis are a deterministic model for determining the base stock
presented in this work, aiming to help on choosing the levels and lead times associated with the lowest cost
most suitable model according to end user’s preferences solution for a supply chain. The stock levels and lead
and purposes. times are determined in such a way as to prevent
stockout and to minimize the amount of obsolete
Keywords: supply chain simulation, Beer Game, inventory at each stock point. In this model, they need
Coloured Petri Nets, constraint programming, models to decide upon two linearly varying demand rates in
analysis, uncertain environment. order to carry out the computation.
Williams (1981) presents seven heuristic
1. INTRODUCTION algorithms for scheduling production and distribution
The Beer Game has a typical supply chain structure operations in an assembly supply chain. The objective
when it is represented as a serially connected inventory of each heuristic is to determine a minimum-cost
management systems chain. The Beer Game application production and/or product distribution schedule that
enables to explore a variety of simple and advanced satisfies the final product demand. However, they fail to
supply management concepts, taking into consideration illuminate the dynamics of the system.
environment uncertainty. Cohen and Lee (1989) present a deterministic,
There are a variety of methods which address mixed integer, non-linear mathematical programming
modelling and analyzing the supply chain. Within them, cost-based model. They use an economic order quantity
simulation has become an important tool for analysis technique to maximize the total after-tax profit for the
and improvement of an entire supply chain operation. manufacturing facilities and distribution centres, but
Different modelling methods that can be used for dynamics are not included.
supply chain analysis may be classified as follows: Modelling supply chains using continuous system
simulation holds great appeal for control theorists. This
1. Analytical modelling: algebraic methods, is because many of the influential characteristics of the
automatic control theory, Petri-Nets, queuing problem can be succinctly expressed in a differential
theory, Markov chains, etc. equation form (Riddals, Bennett and Tipi 2000).
2. Algorithmic modelling: Continuous systems simulation has the advantage of
(a) Continuous Systems Simulation: being a conduit into the frequency domain, which offers
differential equations, difference a framework particularly suited for the study of systems
equations, etc. in which oscillations are a salient attribute, e.g. analysis
of factors having impact on a seasonal, or other demand levels. For more details on DES one could refer to the
fluctuations, amplification as they are passed along the many textbooks, e.g., Law and Kelton (2000), Banks,
chain. Since differential equation produce “smooth” Carson and Nelson (2004), Ho and Cao (1991), etc.
outputs, they are not suited to modelling of all supply Different described methods of mathematical
chains. The system must be considered at an aggregated modelling are suitable for different problems solving
level, in which individual entities in the system and all have their place in the design and management
(products) are not considered. Rather, they are of supply chains. However, the analysis of advantages
aggregated into levels and flow rates. Consequently, and disadvantages of the proposed methods specifies
these methods are not suited for production processes in DES as the more appropriate method for modelling
which each individual entity has an impact on the dynamic systems with a high degree of detailed
fundamental state of the system. For the same reasons, elaboration and stochastic factors, such as supply
continuous systems simulation cannot solve lot sizing chains. For this reason, only models included in the
and job sequencing problems. DES category have been considered in this paper.
Forrester (1961) developed what he called The remainder of this paper is structured as
Industrial Dynamics, which he later extended and follows: next section presents a rough description of the
renamed System Dynamics. He developed a nonlinear Beer Game and introduces the studied models. Section
model of a supply chain using first-order differential 3 presents a discussion about models suitability
equations. He analysed the demand fluctuation according to end user purposes. Finally, some
amplification as it proceed down the multi-echelon conclusions are outlined in the last section.
supply chain, using continuous time models. However,
no sensitivity and cost-based analysis have been carried 2. THE BEER GAME
out on these models, which are solely concerned with The Beer Game is a role-playing simulation developed
the dynamics. Many discrete-event simulation packages at Massachusetts Institute of Technology in the 1960’s
available today provide a more advanced simulation to clarify supply chains’ behaviour (Jarmain 1963). The
capability. Armbuster, Marthaler and Ringhofer (2002) Beer Game model considers a simplified beer supply
modelled high volume production flows using nonlinear chain, consisting of a single retailer, a single wholesaler
hyperbolic partial differential equations, with Little’s which supplies the retailer, a single distributor which
law explicitly built into the formulation. By using the supplies the wholesaler, and a single factory with
developed models, they are able to analyse multiple unlimited raw materials which makes (brews) the beer
products, dispatch polices and control actions. and supplies the distributor (Figure 1).
The emergence of discrete-event systems
simulation (DES) was engendered by the deficiencies of
differential equation approaches to the solution of even
simple man-made problems (Riddals, Bennett and Tipi
2000). Consider, as an example, governing the
behaviour of a series of queues at a supermarket. The
modelling of phenomena such as queue swapping
(when customers jump to shorter queues) or variable
service speed (faster when there are more customers)
would make impossible the application of differential Figure 1: Beer Game’s Supply Chain Structure
equations, as well as any other theoretical approach.
However, such phenomena can easily be incorporated Each component has unlimited storage capacity
into a DES model. and the manufacturer has also unlimited raw materials.
DES has the following two characteristics: (1) it There are a fixed supply lead time and order delay time
represents individual events, e.g., the arrival of an between each participant.
individual customer order; (2) it incorporates Every week, each component in the supply chain
uncertainties, e.g., customer orders arrive at random tries to meet the demand of the downstream participant.
points in time, machines break down at random points Any orders which cannot be fulfilled are recorded as
of time, etc. (Kleijnen 2005). Most systems dynamics backorders. These unmet orders are to be satisfied as
models are non-stochastic, but their behaviour often soon as possible, since no orders can be ignored. At
becomes incomprehensible due to nonlinear feedback each period, each member orders some amount from its
loops. Most econometric models are also based on the upstream supplier. It takes one week for this order to
deterministic, nonlinear differential equations. DES arrive at the supplier. Once the order arrives, the
provides more accurate simulation capabilities against supplier attempts to fill it with its available inventory
above described techniques and so it has been and it takes an additional delay time, commonly set to
considered an important method in supply chain two weeks, before goods arrive to the customer who
modelling. Banks, Buckley and Jain (2002) described a placed the order. Usually, each supply chain component
lot of DES based studies, like commercial packages has no knowledge of the external demand or the orders
developed by IBM for supply chain management and inventory of the other members. Nevertheless, in
simulation for both operational and strategic planning some cases, all components may share information in
order to optimise supply chain’s behaviour (Simchi- The Computerised Beer Game is mainly aimed to
Levi, Kaminsky, and Simch-Levi 2003). education and training on supply chain management.
At each period, each component owns the Although results obtained are equivalent to other
inventory at that facility and goods in transit to the models and the interactive role may be switched off, so
downstream participant. Each location is charged $1 per all participants are controlled by the computer, its
item that it owns as inventory holding cost. In addition, graphical interface significantly slows down its
any backordered item is charged $2 per week. The performance. Therefore, it might not be a good option if
external demand is uncertain and the goal of the retailer, the simulation is aimed to analytical purposes, for
wholesaler, distributor, and factory, is to minimise total which other models may provide the same results faster.
cost, either individually, or for the whole system.
2.2. Coloured Petri Nets Model
2.1. Computerised Beer Game Coloured Petri Nets (CPN) formalism has proven to be
The Computerised Beer Game is a Windows based a successful tool for modelling the characteristics for
program written in C++, developed by Kaminsky and any type of discrete event oriented system. CPN shows
Simchi-Levi (1998), providing an interactive tool for several advantages such as the conciseness of
teaching some supply chain behaviour characteristics. embodying the static structure and the dynamics, the
The Computerised Beer Game follows the original availability of the mathematical analysis techniques as
rules of the Beer Game, with few exceptions aimed to well as its graphical nature (Jensen 1997).
enhance teaching possibilities. The end user can only The Beer Game has been modelled using timed
take one role, usually the distributor, while the Hierarchical Coloured Petri Nets (Jensen 1997)
computer manages all remaining components according following the general scheme presented in Panic,
to the chosen policies (Figure 2). These characteristic Vujosevic and Makajic-Nikolic (2006). The top level of
allows the player focusing on single managerial the model is the whole supply chain represented in the
decisions rather than understanding the whole chain Beer Game, including a customer and the four described
behaviour, for which few information is known. agents: retailer, wholesaler, distributor and
Furthermore, demand in the Computerised Beer Game manufacturer (Figure 5).
may be chosen to be either completely deterministic, as The customer is presented by a place, whose initial
in the original game description, or random, following a marking specifies its demand in time. Retailer,
statistical distribution. wholesaler and distributor are modeled by a Supplier
sub-model (Figure 3). Finally, the manufacturer is
represented by a Manufacturer sub-model (Figure 4).
Furthermore, this hierarchical model allows including
additional suppliers between the customer and the
manufacturer, each of them modelled by an instance of
the Supplier sub-model. This is possible due to in the
Beer Game it is assumed all participants make decisions
according to the same rules. Thus, all agents included in
the supply chain are equal from a modelling
perspective.
Although the CPN model may be extended by
adding additional parameters, only orders, backorders
Figure 2: User Interface of the Computerised Beer and deliveries have been considered. These variables
Game are enough to trace system’s behaviour in order to show
and analyze the bullwhip effect.
However, main changes with respect to the original An instance of the sub-model Supplier has been
beer game are the options to play with global used to represent the retailer, wholesaler and distributor.
information, centralized information and/or shortened According to the order received in place Demand
lead time. When playing the global information (downstream), the current inventory at place
scenario, all information is always known, including Stock and backordered items in place Backorders,
customer demand and inventories. In the centralized a supplier makes its own order. This process is
information version, the player can only take the role of modelled through the instantaneous transition Place
the manufacturer. Because the system is centralized, order and the function order(b,k,n), where
only this component can place orders, while goods are different policies may be used. These policies may be
moved downstream as quickly as possible. As in the implemented in a deterministic way or kept open to
previous option, all information is always available. allow interaction with end users, according to
This option permits to compare centralized and simulation goals. Transition Fulfil demand is used
decentralized policies if costs are correctly adjusted, to model requested amount’s delivery. If the inventory
since no backorders are allowed in the decentralized stores enough goods, the complete demand, including
scenario. Finally, the short lead time version allows last received order and backordered items, is satisfied
reducing the delivery delay from two weeks to one. and the remaining (function rest(n,b)) is kept in the
stock. Otherwise, all available goods are delivered and
the difference is backordered (rest(b,n)). The
duration of transition Fulfil demand is @+2, since
the Beer Game rules establish that deliveries last 2
weeks. The associated guard function ensures this
transition is only fired when an order exists, or there are
backorders to satisfy, and there are goods in stock.
The Manufacturer sub-model is similar to the
Supplier one. In fact, the manufacturer acts as other
suppliers, but deciding what amount to produce in the
following period instead of placing an order to its
upstream agent. Again, this decision is made according
to the current demand, backorders and inventory. The
associated transition is Manufacture, which has
associated a duration of 2 weeks (@+2) as MIT Beer
Game rules state.

Figure 3: Sub-Model Supplier

Figure 4: Sub-Model Manufacturer

Figure 5: Complete Beer Game CPN Model

This model may be implemented in several
platforms supporting CPN simulation, such as CPN DEL1it=DEL2i t-1 (5)
Tools (2010), or using any programming language DEL2it=SHIPi+1 t (6)
(Guasch, Piera, Casanovas and Figueras 2002).
According to implementation details, the model is likely COSTit=1*(INVit+DEL1it+DEL2it)+2*BOit (7)
to be used with different purposes. As an example, all
policies may be implemented so all participants behave The total cost for component i for the whole
in a deterministic way. On the other hand, function simulation period is then calculated trivially:
order may be left in blank for one, some, or all
participants, allowing end users interaction. Thus, an t max
interactive CPN model is implemented, especially TCOSTi = ∑ COSTit (8)
suitable for educational and training purposes, t =1
comparable to the original Beer Game. Taking into
account extra variables may be included in the model, The value of ORDit in (2) is determined according
this CPN model is more likely to be extended according to the chosen policy. Several policies may be
to training preferences and dynamics to be studied. implemented, usually depending on inventory and
demand parameters.
2.3. Constraint Programming Model Rule (3) is only applied to update parameters when
Constraint Programming (CP) is a powerful paradigm there is enough stock to fulfil the current demand. On
for representing and solving a wide range of the other hand, rule (4) is used whenever the component
combinatorial problems. Problems are expressed in i is not able to meet demand requirements. Constraints
terms of three entities: variables, their corresponding (5) and (6) update transportation variables at each
domains and constraints relating them. The problems period.
can then be solved using complete techniques such as Departing from a given initial state and a list of
depth-first search for satisfaction and branch and bound customer’s demand along periods, CP propagation rules
for optimization, or even tailored search methods for determine immediately all remaining variables.
specific problems. Rossi et al. (2006) presents a Therefore, the model is able to provide instantaneously
complete overview of CP modelling techniques, results corresponding to a complete simulation.
algorithms, tools and applications. Moreover, since propagation rules are not
The CP model may be seen as a specific unidirectional, the model may provide a mechanism to
implementation of the CPN model described in the infer other participants’ policies, inventory bounds and
previous section. Constraints among variables are even final customer’s demand. With this goal, initial
defined as a set of rules, relating each component’s information concerning the evolution of own variables
variables with its upstream/downstream participant and should be provided, instead of final customer’s demand.
their values at each period. So, the model may reconstruct demand patterns from
Four sets of variables have been defined, one for other participants, even for the final customer, and
each component i=[1..4] (1: retailer, 2: wholesaler, 3: provide a good mechanism for analysis and a first step
distributor, 4: manufacturer) and the final customer to get a reliable forecasting tool.
being i=0. Periods are denoted by the variable The CP model, implemented using the CP platform
t=[1..tmax], where tmax is usually set to 50 weeks. Thus, ECLiPSe (Apt and Wallace 2007), is instantaneous for
for component i at period t, variables defined are: INVit common simulation periods. Since propagation is very
is the current inventory, DEL1it and DEL2it are the fast, the model is especially suitable for analysis
goods in transit, BOit is the number of backordered purposes. Furthermore, it is likely to be parallelized, so
items, DEMit is the demand to be satisfied, ORDit is the different scenarios may be defined and run
order placed by the component, SHIPit is the amount simultaneously to get different data. Nevertheless, the
shipped by the component in the current period and model provides a low level of interaction, so it might
COSTit is the associated cost to this turn. Variables are not be a good option for decision making training.
related according to the following rules:
2.4. Arena
DEM1t=ORD0 t (1) Simulation techniques are used when analytical solving
DEMit=ORDi-1 t-1
(2) is impossible. Most of all analytical approaches do not
succeed analysing complex, dynamic systems like
INVi t-1 + DEL1i t-1 • DEMit + BOi t-1
→ (3) supply chains. For this reason, the Beer Game has been
INVit=INVi t-1+DEL1i t-1–SHIPit modelled using DES software Arena by Rockwell
SHIPit=DEMit+BOi t-1 Automation (2010). Arena software is effective when
BOit=0 analysing complex, medium to large-scale projects,
involving highly sensitive changes related to supply
INVi t-1 + DEL1i t-1 < DEMit + BOi t-1
→ (4) chain, manufacturing, processes, logistics, distribution,
INVit=0 warehousing and service systems. Arena proposes an
SHIPit=INVi t-1+DEL1i t-1
BOit=BOi t-1+DEMit-SHIPit event-oriented modelling approach and consists of
"libraries" of modelling objects that make it As Arena’s simulation engine do not always
significantly easier and faster to develop models. Arena process the events in this order (Kelton and Sadowski
exploits two Windows technologies that are designed to 2002), a procedure has been developed to guarantee that
enhance the integration of desktop applications. The events are processed in the mentioned sequence. Wait
first, ActiveX Automation allows applications to control and Signal blocks form implementation’s basis of
each other and themselves via a programming interface. this procedure in Arena.
The second technology addresses the programming Various experiments may be performed with the
interface issues to a Visual Basic programming created model and achieved results may be analysed by
environment. These two technologies work together to the Output Analyzer. This is a component of Arena that
allow Arena integrating with other programs that provides an easy-to-use interface, simplifying data
support ActiveX Automation, e.g., Excel, AutoCAD, or analysis and allowing viewing and analysing simulation
Visio. data quickly and easily.
The Beer Game supply chain and its inventory The Arena model is suitable for research purposes,
control systems have been implemented combining since it incorporates both dynamic and stochastic nature
Arena software and Visual Basic for Applications of the supply chain operations and the simulation
(VBA). The model logic is represented comprehensibly execution speed without animation is quite fast. The
in the Arena flowchart-style environment, while the analysis of system’s behaviour under specified
more complex calculation algorithms are programmed conditions may be easily performed and allows
in VBA. The traditional four-stage Beer Game structure foreseeing thousands of situations which could result
has been modelled in combination with two information from supply chain operations. Each of the supply chain
sharing strategies (centralised and decentralised) and stages is modelled as a separate module and it is
two inventory control policies (s-S and Stock-To- possible to change both supply chain structure (e.g., add
Demand). or remove a stage) and supply chain management
In accordance with specific features of the Beer concepts (e.g., centralized or decentralized information
Game, the proper event processing schedule sharing strategy).
implementation in Arena environment is shown in However the model is not suited for training and
Figure 6. If several events are scheduled to occur at a education due to low level of interaction with the end
certain supply chain stage at the same simulation time, user. To modify the model parameters, end user should
there is a fixed order in which the events should be have previous experience on working with Arena
processed: software.

1. order or backorder arrival from upstream stage 2.5. Excel

(stock replenishment); An alternative to DES specific software is developing a
2. fulfilling of backorders (only if an order has supply chain simulator implemented in a general-
arrived); purpose high-level programming language, e.g., C++,
3. new demand fulfilling. Java or VBA. Programming languages are mainly used
for simulation in order to avoid additional expenses of
commercial software purchasing and maintenance.
The user interface of the Beer Game supply chain
model is developed by means of Microsoft Excel
spreadsheets (Figure 7), which is widely applied and
easy accessible software. The programming logic is
implemented using VBA (Ternovoy 2004).

Figure 6: Submodel of Order Shipment to Wholesaler Figure 7: Supply Cain Model User Interface
Supply chain’s control variables’ values, as well as Table 1 summarizes characteristics for all
initial data, are defined by the user using a MS Excel described models in the previous section. Even though
interface, and then the simulator is run for the specified different parameters might be defined, only those
number of periods by means of a VBA procedure relevant for purposes considered in this paper have been
(Figure 8). included, i.e. interactivity, scalability, execution speed
and the capability of inferring demand patterns from a
data set. Interactivity is a characteristic of
Computerised, CPN and Excel models. In all of them,
different parameters can be chosen or modified on
runtime, or the model is likely to be modified easily to
include some level of interaction with the end user. On
the other hand, the CP model is completely
deterministic and results are only derived from the
initial data and set-up. Arena model requires some
particular skills in order to permit some interaction with
the end user. Models’ modularity determines their
capability of being scaled to represent larger systems. In
this sense, Arena, Excel, CPN and CP models are
clearly modular, since different components are defined
separately and may be concatenated with little changes.
The computerised Beer Game is proprietary software
Figure 8: Simulation Parameters and so the system it represents cannot be modified.
According to the simulation speed, CP and Arena
The simulation procedure Next_n runs the model models are clearly faster than other models. This
in accordance with defined user settings. First, constants characteristic makes them especially efficient when
and variables are defined and information is read from running a set of simulations with different parameters,
the data files. Then, the simulation of the processes as demanded for analysis purposes. Finally, only the CP
scheduled is performed and repeated four times for all model allows performing simulations with different
supply chain’s stages, i.e. once for each component. initial data than the final customer demand and
Events management is performed in accordance with participants’ policies. This characteristic provides a
the Beer Game structure. Eventually, simulation results mechanism especially important from a forecasting
appear in separate spreadsheets showing different types perspective. As mentioned in the CP model description,
of charts, histograms and tables. it also permits reconstructing demand patterns and
The functional possibilities of the Excel simulation inferring other participants’ policies and inventory
model are quite similar to the Arena model, but since bounds. Therefore, this characteristic might be an
the special simulation software is not needed and interesting tool for analysis, either for research or
experiments are easily configured, it can be used for management.
educational and training purposes. However, the speed
of executing a simulation run in Excel is quite slow and Table 2: Models’ Suitability According to Simulation
the number of simulated periods is restricted, so Purposes
research tasks are difficult by using this model. Training / Research
Model Strategies
Education Analysis
3. MODELS SUITABILITY develop.
Although all described models yield the same results, Computerised X
choosing one or another relies on simulation’s goals. CPN X X
Each model’s characteristics make it more suitable CP X X
according to pursued purposes: research or Arena X
training/education. However, all models may be used to Excel X
simulate the Beer Game regardless of their
characteristics, being only a matter of efficiency which All parameters related in Table 1 characterize
one is best suited to a specific simulation goal. models’ performance and adaptability, and help on
choosing one or another depending on simulation
Table 1: Models’ Characteristics Summary purposes. For example, if the simulation is training or
Infer education oriented, requested characteristics mainly
Model Interactive Scalable Speed
Policies include a high interactivity, in order to permit the end
Comput. X user making his own decisions and introducing them
CPN X X into the system. These decisions could be programmed
CP X X X in advance for all models, so not interactivity is allowed
Arena X X at all, but it would reduce the teaching or training
Excel X X experience to a single analysis after the simulation ends.
For this reason, interactive models among described This way, thousands of different scenarios could be
ones, i.e. Computerised Beer Game, CPN and Excel, are simulated sequentially, or even in a parallelized
considered to be best suited for management training or environment, getting results in reasonable times. On the
education. On the other hand, research tasks demand other hand, there is a lack of interactivity with the end
another kind of characteristics, especially related to user that makes it difficult to be used with real-time
execution speed. Usually, separate sets of simulations decision making training.
are run combining different demand patterns and The Arena model incorporates advantages from
policies, in order to analyse system’s behaviour. With using specific simulation software, such as the
this goal in mind, it is more appropriate to choose a capability of including both dynamic and stochastic
model which allows running a complete simulation in a nature of supply chains operations. Since it is able to
low time, such as CP or Arena models. Although the provide results in a quite reasonable time, this model is
execution time depends on the implementation platform suitable to be used for research purposes. However,
and computer’s characteristics, both models have using Arena on developing and modifying supply chain
demonstrated being fast enough for research tasks. models requires particular skills on working with
Another desirable model characteristic is scalability, simulation software environment.
since it allows modifying system’s dimensions with Using Excel for developing a simulation model
analytical purposes. In this sense, both CPN and CP becomes an alternative to using Arena. Both models
models permit changing the number of components possibilities are similar, having the advantage that no
easily while ensuring a complete integration of all special simulation software is needed. Thus, Excel
variables. Finally, the capability of the CP model to model may also be used for educational and training
infer other components’ policies, demand patterns and purposes. However, its simulation speed is quite low
inventory bounds, combined with its low execution compared to Arena, so research tasks are limited when
time, make it especially suitable for strategies and using this model.
policies development, as well as a first step for Described models’ main purposes are clearly
developing forecasting tools. Table 2 presents which determined by parameters analysed in section 3, among
models best fulfil different simulation goals. others. As an example, for training simulations is highly
Nevertheless, conclusions presented in Table 2 are to be desirable to run a model which allows interacting with
considered more a guide than a rule, since all models the end user. Thus, he can check almost instantly
may be used for all selected purposes according to end consequences derived from his decisions. On the
user preferences, as mentioned. contrary, research tasks often require a high execution
speed, so different scenarios may be simulated and
4. CONCLUSIONS analysed in low times. Scalability is another desirable
The Beer Game has a sequential supply chain structure characteristic, even though it is not critical due to Beer
that permits exploring many supply management Game models simplicity. It should be remarked a
concepts. In the present paper, four discrete-event characteristic that only the CP model possesses: inverse
simulation models representing the Beer Game have deduction and reconstruction of policies and inventory
been described. Some of their characteristics have also bounds. CP paradigm allows this characteristic, since
been outlined and related to their suitability according constraints are not unidirectional, unlike rules included
to simulation purposes. This analysis may form a in other models.
helpful basis for choosing the most suitable model Finally, it should be remarked that all four
according to end user’s preferences and purposes. presented models are simulation oriented. Simulation
The Computerised Beer Game is mainly aimed to models are the so-called input-output models, i.e., they
education and supply chain management training. It yield the output of the system for a given input.
allows a high level of interaction with the end user, Therefore, simulation models are “run” rather than
although the interactive role may also be switched off “solved” (Hurrion 1986). For this reason, no one of
aiming to increase the simulation speed. However, its presented models may be used for optimization
graphical interface significantly slows down model’s purposes without significant changes. In any case,
performance, so it might not be the best option if the system may be studied and optimised, but depending
simulation has analysis purposes. entirely on the end user. Among models described in
CPN provide a mechanism to specify a conceptual the present paper, only the CP model is suitable to be
model, likely to be implemented either on CPN easily modified to include an objective function in order
simulation platforms or by using high-level to be optimised. However, supply chains nature would
programming languages. Thus, it possesses a set of induce to explore a huge search space, characteristic of
characteristics that permit using it for educational and NP problems. Tackling this kind of problem would
training purposes, as well as for analysis and research require using complex algorithms, i.e. heuristics and
tasks. metaheuristics, combined with simulation and other
The CP model may be considered as an operational research techniques.
implementation of the proposed CPN model. Since
constraint propagation is very fast, it provides almost
instantaneous results for common simulation periods.
REFERENCES Kleijnen, J.P.C., 2005. Supply chain simulation tools
Apt, K., and Wallace, M.G., 2007. Constraint and techniques: a survey. International Journal of
Programming using ECLiPSe. Cambridge: Simulation & Process Modelling, 1(1-2): 82-89.
Cambridge University Press. Law, A.M., and Kelton, W.D., 2000. Simulation
Armbuster, D., Marthaler, D., and Ringhofer, C., 2002. modeling and analysis. International Series in
Efficient simulations of supply chains. Industrial Engineering and Management Science.
Proceedings of the 2002 Winter Simulation Singapore: McGraw-Hill.
Conference, 1345-1348. December 8-11, San Panic, B., Vujosevic, M., and Makajic-Nikolic, D.,
Diego (USA). 2006. A Supply Chain Simulation Model Based on
Banks, J., Buckley, S., and Jain, S., 2002. Panel session: Hierarchical Coloured Petri Nets. Proceedings of
opportunities for simulation in supply chain the International Scientific Days –
management. Proceedings of the 2002 Winter Competitiveness in the EU – Challenge for the V4
Simulation Conference, 1652-1658. December 8- Countries, 1431-1436. May 17-18, Nitra
11, San Diego (USA). (Slovakia).
Banks, J., Carson, J.S., and Nelson, B.L., 2004. Riddalls, C.E., Bennett, S., and Tipi, N.S., 2000.
Discrete-Event System Simulation. New Jersey: Modelling the dynamic of supply chain.
Prentice Hall. International Journal of System Science, 31(8):
Bessiere, C., 2006. Constraint propagation. In: Rossi, 969-976.
F., van Beek, P. and Walsh, T., eds. Handbook of Rockwell Automation, 2010. Arena Simulation.
Constraint Programming. Amsterdam: Elsevier, Available from: www.arenasimulation.com.
29-83. Rossi, F., van Beek, P., and Walsh, T. (eds.), 2006.
Cohen, M.A., and Lee, H.L., 1989. Resource Handbook of Constraint Programming.
deployment analysis of global manufacturing and Amsterdam: Elsevier.
distribution networks. Journal of Manufacturing Simchi-Levi, D., Kaminsky, P., and Simchi-Levi, E.,
and Operations Management, 2: 81-104. 2003. Designing and Managing the Supply Chain:
CPN Tools, 2010. CPN Tools, Computer Tool for Concepts, Strategies and Case Studies. New York:
Coloured Petri Nets. Available from: McGraw-Hill.
wiki.daimi.au.dk/cpntools. Ternovoy, V., 2004. Simulation-Based Analysis of the
Forrester, J.W., 1961. Industrial Dynamics. New York: Supply Chain Operation. Master thesis (supervisor
MIT Press and Wiley. Petuhova, J.). Riga Technical University.
Guasch, A., Piera, M.A., Casanovas, J. and Figueras, J. Williams, J.F., 1981. Heuristic techniques for
(eds.), 2002. Simulación de Sistemes Orientados a simultaneous scheduling of production and
Eventos Discretos. In: Modelado y Simulación. distribution in multi-echelon structures: theory and
Aplicación a Procesos Logísticos de Fabricación y empirical comparisons. Management Science,
Servicios. Barcelona: Edicions UPC. 27(3): 336-352.
Ho, Y.C., and Cao, X.R., 1991. Perturbation Analysis
of Discrete Event Dynamic. Boston: Kluwer ACKNOWLEDGMENTS
Academic Publisher. This work has been supported by the Agència de Gestió
Hurrion, R.D., 1986. Simulation: Applications in d'Ajuts Universitaris i de Recerca (AGAUR) of the
manufacturing - International trends in Generalitat de Catalunya (project 2009 SGR 629).
manufacturing technology. Bedford: Springer-
Verlag.
Ishii, K., Takahashi, K., and Muramatsu, R., 1988. AUTHORS BIOGRAPHY
Integrated production, inventory and distribution Daniel Guimarans is a PhD Student and Assistant
systems. International Journal of Production Teacher at the Telecommunication and Systems
Research, 26(3): 473-482. Engineering Department at the Universitat Autònoma
Jarmain, W.E., 1963. Problems in Industrial Dynamics. de Barcelona. He is an active member of the Logisim
Cambridge: MIT Press. Research Group. His research interests are constraint
Jensen, K., 1997. Coloured Petri Nets. Basic Concepts, programming, simulation and hybridisation of different
Analysis Methods and Practical Use. Volumes 1, techniques to solve industrial combinatorial problems.
2, 3. Berlin: Springer-Verlag.
Kaminsky, P., and Simchi-Levi, D., 1998. A new Julija Petuhova is Doctor of Engineering, lecturer of
computerized beer game: a tool for teaching the the Institute of Information Technology at Riga
value of integrated supply chain management. In: Technical University. Her research interests include
Hau Lee and Shu Ming Ng, eds. Global Supply simulation methodology of logistics systems, supply
Chain and Technology Management. Miami: The chain dynamics, practical applications of simulation
Production and Operations Management Society, modelling, training and education via simulation-based
216-225. business games. She is a member of the Latvian
Kelton, W.D., and Sadowski, R.P., 2002. Simulation Simulation Society and has a wide experience in
with Arena. New York: McGraw-Hill.
performing research projects in the simulation area at
national level.

Prof. Yuri Merkuryev is Habilitated Doctor of

Engineering, Professor of the Institute of Information
Technology at Riga Technical University, Head of the
Department of Modelling and Simulation. His
professional interests include methodologies and
practical implementation of discrete-event simulation,
supply chain modelling and management, and education
in the areas of modelling, simulation and logistics
management. He is Programme Director of the Master-
level curriculum “Industrial Logistics Management” at
Riga Technical University. Prof Merkuryev has a wide
experience in performing research and educational
projects in the simulation area, at both national and
European levels, and regularly participates in organising
international conferences in the simulation area. He has
promoted 5 PhD theses and authors about 250 scientific
publications, including 5 books and numerous journal
and conference papers, as well as a text-book in
Logistics Information Systems. He is also co-editor of a
collection of simulation-based case studies in logistics,
published in 2009 by Springer-Verlag. Prof. Merkuryev
is a Board member of the European Council of the
Society for Modeling and Simulation International,
President of the Latvian Simulation Society, and Board
member of the Latvian Transport Development and
Education Association.

Juan José Ramos is an Associated Professor at the

Telecommunication and Systems Engineering
Department at the Universitat Autònoma de Barcelona.
He is coordinator and an active member of the Logisim
Research Group. His research interests are modelling,
simulation and optimisation of logistics problems.

View publication stats