Omega 33 (2005) 419 423

www.elsevier.com/locate/omega

Information sharing in supply chains

P. Fiala
University of Economics, Prague, Czech Republic
Accepted 29 June 2004
Available online 15 September 2004

Abstract
Supply chain is dened as a system of suppliers, manufacturers, distributors, retailers and customers where material, nancial
and information ows connect participants in both directions. Most supply chains are composed of independent agents with
individual preferences. It is expected that no single agent has the power to optimise the supply chain. Supply chain management
is now seen as a governing element in strategy and as an effective way of creating value for customers. The so-called bullwhip
effect, describing growing variation upstream in a supply chain, is probably the most famous demonstration that decentralised
decision making can lead to poor supply chain performance. Information asymmetry is one of the most powerful sources of
the bullwhip effect. Information sharing of customer demand has an impact on the bullwhip effect. Information technology
has lead to centralised information, shorter lead times and smaller batch sizes. The analysis of causes of the bullwhip effect
has lead to suggestions for reducing the bullwhip effect in supply chains by strategic partnership. Supply chain partnership
leads to increased information ows, reduced uncertainty, and a more protable supply chain. The cooperation is based on
contacts and formal agreements. Information exchange is very important issue for coordinating actions of units. New business
practices and information technology make the coordination even closer. Information sharing and strategic partnerships of
units can be modelled by different network structures.
2004 Elsevier Ltd. All rights reserved.
Keywords: Supply chain; System dynamics; Bullwhip effect; Information sharing; Cooperation

1. Supply chain management

Supply chain management has generated a substantial
amount of interest both among managers and researchers.
Supply chain management is now seen as a governing element in strategy and as an effective way of creating value
for customers. Supply chain management benets from a
variety of concepts that were developed in several different
disciplines as marketing, information systems, economics,
system dynamics, logistics, operations management, and operations research. There are many concepts and strategies
applied in designing and managing supply chains [1]. The
expanding importance of supply chain integration presents
a challenge to research to focus more attention on supply
E-mail address: pala@vse.cz (P. Fiala).
0305-0483/$ - see front matter 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.omega.2004.07.006

chain modelling [2]. A structure of supply chains is composed from potential suppliers, producers, distributors, retailers and customers etc. The units are interconnected by
material, nancial, information and decisional ows. Most
supply chains are composed of independent units with individual preferences. Each unit will attempt to optimise his
own preference. Behaviour that is locally efcient can be
inefcient from a global point of view. In supply chain behaviour are many inefciencies. An increasing number of
companies in the world subscribe to the idea that developing long-term coordination and cooperation can signicantly
improve the efciency of supply chains and provide a way
to ensure competitive advantage.
The overall business environment is becoming increasingly dynamic. Demand and supply for custom products
can be very dynamic. Supply chains operate in a network

420

P. Fiala / Omega 33 (2005) 419 423

environment. Dynamic information and decision-making

models are called to accommodate these changes and uncertainties. There are some approaches to model and analyse
the supply chain dynamics. AND/OR networks are capable
of capture a variety of activities and decisions and can be
used by decision-makers to evaluate the efciency of current operations and to coordinate supply chain activities.
Dynamic models use concepts of state variables, ows, and
feedback processes and can be developed and simulated
with help of STELLA software, where graphical tools are
utilised. Dynamic models of supply chains try to reect
changes in real or simulated time and take into account that
the network model components are constantly evolving.
An information asymmetry is a source of inefciency
in supply. Information exchange is a very important issue
for coordinating actions of units. The proposed modelling
framework is composed from three inter-related network
structures: a production net, Petri net and a neural net. The
production net captures information ows among agents.
The Petri net is used to coordinate asynchronous events of
different units in network production systems and to model
negotiation process. The neural net serves as an instrument
for inductive learning of negotiation strategies.
The expected result is a mutually benecial, win-win partnership that creates a synergistic supply chain in which the
entire chain is more effective than the sum of its individual
parts. Supply chain partnership leads to increased information ows, reduced uncertainty, and a more protable supply
chain. The ultimate customer will receive a higher quality,
cost-effective product in a shorter amount of time.

2. System dynamics
System dynamics is concerned with problem solving in
living systems [36]. It links together hard control theory
with soft system theory. System dynamics needs relevant
tools from both ends of the systems spectrum. If the possible
causal factors are identied and their respective contribution to the overall dynamics are quantitatively measured and
benchmarked, then it would be conducive to performance
improvement by eliminating or reducing the relevant dynamics. The performance improvement is also predictable
with the help of the dynamics benchmark. Systems of information feedback control are fundamental to all systems.
Feedback theory explains how decisions, delays and predictions can produce either good control or dramatically unstable operation.
The supply chain dynamics lead to the increase in the
cost of the units and the whole chain. A feedback control
system causes a decision, which in turn affects the original
environment. In supply chains, orders and inventory levels
lead to manufacturing decisions that ll orders and correct
inventories. As a consequence of using system dynamics
in supply chain redesign, we are able to generate added
insight into system dynamic behaviour and particularly into

underlying causal relationships. This new knowledge can be

exploited in the improved design, robustness and operating
effectiveness of such systems.
A multilevel network model was proposed [7]. The model
consists of: the material network, the informational network,
and the nancial network. The multiple decision-makers use
multiple criteria as quantity, time and cost. The efcient
frontier of solutions can be identied. This network model
is appropriate for analysing system dynamics. It can be formulated as a broad class of dynamic supply network problems. Dynamic behaviour of orders, inventories, prices and
costs at different stages of supply network can be analysed.
The so-called bullwhip effect [5,2], describing growing
variation upstream in a supply chain, is probably the most
famous demonstration of system dynamics in supply chains.
The basic phenomenon is not new and has been recognised
by Forrester [4]. There are some known causes [5,2], of the
bullwhip effect: information asymmetry, demand forecasting, lead-times, batch ordering, supply shortages and price
variations. Information sharing of customer demand has an
impact on the bullwhip effect.
The analyses of causes and suggestions for reducing the
bullwhip effect in supply chains are challenges to modelling
techniques. We consider a k-stages supply chain. The customer demands are independent and identically distributed
random variables. The last stage observes customer demand
D and places an order q to previous stage. All stages place
orders to the previous stage in the chain. The orders are received with lead-times Li between stages i and i + 1. The
stages use the moving average forecast model with p observations. To quantify increase in variability, it is necessary to
determine the variance of orders q k relative to the variance
of demands D.
In the case of decentralised information the variance increase is multiplicative at each stage of the supply chain


k
2L2i
Var(q k )

2Li


+ 2 .
1+
Var(D)
p
p
i=1

In the case of centralised information, i.e. the last stage provides every stage of the supply chain with complete information on customer demand, the variance increase is additive:


2( ki=1 Li ) 2( ki=1 Li )2
Var(q k )

1 +
+
.
Var(D)
p
p2
The centralised solution can be used as a benchmark, but
the bullwhip effect is not completely eliminated.
3. Modelling of system dynamics
The structure of supply chains and relations in among
units can be modelled by different types of networks.
AND/OR networks can be applied for modelling exible
and dynamic supply chains [8]. The approach follows an
activity on arc representation where each arc corresponds

P. Fiala / Omega 33 (2005) 419 423

to a particular supply chain activity. Each activity has multiple performance criteria. Nodes represent completion of activities and establish precedent constraints among activities.
The initial suppliers without predecessors and end customers
without successors are represented by nodes displayed as
circles. Two types of nodes are dened to specifying prior
activities. AND nodes are nodes for which all the activities
must be accomplished before the outgoing activities can begin. OR nodes require that at least one of the incoming activities be nished before the outgoing activities can begin.
The STELLA software is one of several computer applications created to implement concepts of system dynamics
[9]. It combines together the strengths of an iconographic
programming style and the speed and versatility of computers. The instrument is very appropriate to proposed modelling framework for dynamic multilevel supply network.
The STELLA language consists of four basic building
and one space-saving tool. The four building blocks are:
Stock, represent something that accumulates.
Flow, activity that changes magnitude of stock.
Converter, modies an activity.
Connector, transmits inputs and information.
Fig. 1 shows the icons of building blocks. This approach
enables to model and to solve a broad class of dynamic
problems. Differential equation can be used for modelling
of system dynamics. STELLA software offers the numerical

CONVERTER

FLOW

STOCK
CONNECTOR

Fig. 1. Building blocks of STELLA language.

421

techniques (Ellers method, RungeKutta-2 and Runge

Kutta-4 methods) to solve the model equations.
STELLA software contains many built-in functions
that can facilitate dynamic modelling of supply networks
(Fig. 2). There are some examples of instruments for proposed modelling approach;
AND/OR to modelling of AND/OR network environment,
DELAY to modelling of lead-times,
DT time step,
FORCST forecasts demand in stages of supply chain,
RANDOM generates random customer demand.
As an example of dynamic problem, a stochastic inventory problem can be analysed with the nite time horizon.
AND/OR supply network consists of a structure of suppliers, different production modes, an assembly of components
and production of an end product to a customer. We can
describe the behaviour of the network decision-makers and
propose a dynamic system that captures the adjustments of
the commodity shipments and the prices over space and
time. The bullwhip effect can be demonstrated by comparison of random customer demand and orders in different
stages of the supply network by decentralised information.
Centralised information of customer demand can reduce the
bullwhip effect.
4. Modelling of cooperation in supply chains
The layers of producers and retailers compete in a
non-cooperative way, but the partners in individual supply
chains can prot from cooperative decision making. The
strategic partnership means cooperation and coordination of
actions through the supply chain. The strategic partnerships
change material, nancial and information ows among

Fig. 2. Supply network by STELLA.

422

P. Fiala / Omega 33 (2005) 419 423

participants in the supply chain. The way of information

sharing is changed by information centralising using information technology. The expected result is mutually benecial, winwin partnership that creates a synergistic supply
chain in which the entire chain is more effective than the
sum of its individual parts. Supply chain partnership leads
to increased information ows, reduced uncertainty, and a
more protable supply chain.
The general suppliercustomer relations in supply chain
can be taken as centralised or decentralised [7]. The partnership relations are based on supply contracts. Contracts provide a means for bringing the decentralised solution to the
centralised solution. Contracts also facilitate long-term partnership by delineating mutual concessions that favour the
persistence of the relationship, as well as specifying penalties for non-cooperative behaviour. The contracts are evaluated by multiple criteria as price, quantity, costs, time and
quality. There are different approaches to modelling multicriteria negotiation processes to reach a consensus among
partners. A cooperative decision making requires free communication among agents and gives synergical effects in a
conict resolution. The basic trend in the cooperative decision making is to transform a possible conict to a joint
problem.
Some basic ideas of formal approaches of problem solving can be introduced to cooperative decision making. There
are two aspects of the problem solvingrepresentation and
searching. The state space representation introduces the concepts of states and operators. An operator transforms one
state into another state. A solution could be obtained by a
search process that rst applies operators to the initial state
to produce new states and so on, until the goal state is produced. Communication between producers and retailers can
be provided through information sharing (see Fig. 3).
The proposed model is a discrete dynamic model and the
cooperation of units is based on contracts and formal agreements achieved in negotiation process. We propose a two
phases interactive approach for solving cooperative decision making problems [10]:
1. Finding the ideal solution for individual agents.
2. Finding a consensus for all the agents.
In the rst phase every decision maker search the ideal alternative by the assertivity principle. In the second phase a
consensus could be obtained by the search process and the
principle of cooperativeness is applied. The heuristic information for the decision making unit is the distance between
his proposal and the opponents proposal.
The negotiation process modelling in general is a complex
problem based on several kernel ideas. The framework [7]
of the proposed discrete dynamic model is separated to three
parts
deterministic part,
logical part,
stochastic part.

Producer i
- information
- tools

Shared
information
Producers i
Retailers j
problem
problem
Joint problem
representation representation representation

Retailer j
- information
- tools

Coordinator
- information
- tools

Fig. 3. Communication through information sharing.

According to these three parts the modelling framework is

composed of three inter-related network structures
ow network,
Petri network,
neural network.
The ow network represents the supply chain network with
information, material and nancial ows between partners.
The Petri network is used to coordinate asynchronous events
of different units in the supply chain and to model negotiation process. The neural network serves as an instrument
for inductive learning of negotiation strategies.
Producers and retailers in individual supply chains negotiate contracts about prices, quantities, costs, time, etc. The
double marginalisation problem can be generalised and producers and retailers negotiate the prot-sharing contracts.

5. Conclusions
Supply chain management has generated a substantial
amount of interest both among managers and researchers.
The interest has also been fuelled by the growth in the development and application of e-business technologies. Ebusiness is associated with business models and practices
enabling continuous improvements in supply chains. Supply
chain management is more and more affected by network
and dynamic business environment and by information and
communication technologies. The network economy is characterised by massive global connectivity relationship among
economic subjects.
The paper is devoted to modelling of supply chain dynamics. In the model some important features of this environment are established. The combination of network structure
modelling and simulation of dynamic behaviour of units in
supply chain can be a powerful instrument of performance
analysis of supply chains. Multicriteria analysis of supply
chain performance includes criteria such as quantity, time,
cost and prot. Simulation approach by STELLA software
is an appropriate tool for prediction of real supply chain
situation. The partners in supply chains can prot from
cooperative decision making. The combination of noncooperative and cooperative behaviour of network users is
more realistic.

P. Fiala / Omega 33 (2005) 419 423

The proposed modelling framework is composed from

three inter-related network structures: a production net, a
Petri net and a neural net. The production net captures information ows among agents. Petri net is used to coordinate
asynchronous events of different units in network production systems and to model negotiation process. The neural
net serves as an instrument for inductive learning of negotiation strategies.
Acknowledgements
The research project was supported by Grant No.
402/01/0771 from the Grant Agency of Czech Republic
Modeling of Supply Chains and CEZ: J 18/98: 311401001
from the University of Economics Models and Methods
for Economic Decisions.
References
[1] Simchi-Levi D, Kaminsky P, Simchi-Levi E. Designing and
managing the supply chain: concepts strategies and case
studies. Homewood, IL/New York: Irwin/McGraw-Hill; 1999.

423

[2] Tayur S, Ganeshan R, Magazine M. Quantitative models for

supply chain management. Dordrecht: Kluwer; 1999.
[3] de Souza R, Song Z, Chaoyang L. Supply chain
dynamics and optimization. Integrated Manufacturing Systems
2000;11(5):34864.
[4] Forrester JW. Industrial dynamics. Cambridge, MA: MIT
Press; 1961.
[5] Lee HL, Padmanabhan V, Whang S. Information distortion
in a supply chain: the bullwhip effect. Management Science
1997;43:54658.
[6] Swaminathan J, Smith S, Sadeh N. Modeling supply
chain dynamics: a multiagent approach. Decision Sciences
1998;29(3):60732.
[7] Fiala P. Modeling of relations in supply chains, Vision: The
Journal of Business Perspective. Special Issue on Supply
Chain Management 2003;7:12731.
[8] Zeng DD. Managing exibility for inter-organizational
electronic commerce. Electronic Commerce Research
2001;1(12):3351.
[9] Ruth M, Hannon B. Modeling dynamic economic systems.
Berlin: Springer; 1997.
[10] Fiala P. Models of cooperative decision making. In: Gal T,
Fandel G., editors. Multiple criteria decision making. Berlin:
Springer; 1997.