AGENT-BASED MODELING AND SIMULATION OF RESOURCE
ALLOCATION IN ENGINEERING CHANGE MANAGEMENT
Bochao Wang and Young B. Moon
Departement of Mechanical and Aerospace Engineering, Syracuse University, Syracuse, NY 13244, U.S.A.
Keywords: Engineering Change Management, Agent-based Modeling and Simulation, Resource Allocation.
Abstract: An engineering change (EC) refers to a modification of products and components including purchased parts
or even supplies after product design is finished and released to the market. While any company involved in
product development would have to deal with engineering changes, the area of engineering change
management hasn't received much attention from the research community. It is partly because of its
complexity and lack of appropriate research tools. In this paper, we present preliminary research results of
modeling the engineering change management (ECM) process using an agent-based modeling and
simulation technique. The aim of the research reported in this paper is to study optimal strategies of
resource allocation for a company when it is dealing with two kinds of ECs: "necessary ECs" and
"initialized ECs." We discuss results from these simulation models to illustrate some insights of ECM, and
present several research directions from these results.
1 INTRODUCTION
Any company involved in product development
would have to deal with engineering changes. An
engineering change (EC) refers to a modification of
products and components including purchased parts
or even supplies after product design is finished and
released to the market (Lee 2006), (Clark &
Fujimoto, 1991), (Huang, Yee & Mak, 2003),
(Terwiesch & Loch, 1999), (Chen 2002, DEC et al.
1998). ECs may be initiated by customers, suppliers,
or the company itself.
While EC management is a complex task, the
increasing market competition forces the companies
to take a more pro-active role in handling the
engineering changes. In a sense, any engineering
change is a disruption to a normal operation. And it
may impact several functions across a company.
However, an effective and efficient management of
engineering changes can bring significant benefits to
company’s competitiveness (Rukta, 2006) by
satisfying its customers better and further improving
its products.
Despite of its importance, the area of engineering
change management hasn't received much attention
from the research community. Notable exceptions
are the works of Nadia (Nadia, 2006) and Terwisch
(Terwiesch, 1999). Nadia studied engineering
change orders (ECOs) thoroughly and even
identified key contributors to long ECO lead times
with improvement strategies advices. Terwiesch
compared the behavior of two methods of managing
an engineering change request (ECR) process.
The aim of the research reported in this paper is
to study optimal strategies of resource allocation for
a company when it is dealing with two kinds of ECs:
necessary ECs and initialized ECs. Necessary ECs
refer to those ECs dealing with must-to-do changes
to address problems such as technical problem,
manufacturing process or design fault. Initialized
ECs may arise from introducing new technology to
match competitors, to take the lead, to simplify or
improve manufacturing processes, or to
accommodate customers’ proposal. So the latter is
not mandatory but may bring potential benefits to
the company.
2 THE RESEARCH
METHODOLOGY
2.1 ABMS
ABMS (Agent-based Modeling and Simulation) is a
computational model for simulating the actions and
interactions of autonomous individuals in a network,
281
Moon Y. and Wang B. (2009).
AGENT-BASED MODELING AND SIMULATION OF RESOURCE ALLOCATION IN ENGINEERING CHANGE MANAGEMENT.
In Proceedings of the 11th International Conference on Enterprise Information Systems - Artificial Intelligence and Decision Support Systems, pages
281-284
DOI: 10.5220/0001852302810284
Copyright
c
 SciTePress

with a view to assessing their effects on the system
as a whole. It is a relatively new approach to
modeling complex systems (Macal, 2005) in order to
support decision-making (Nilsson, 2006) and obtain
deeper understanding of intrinsic regularity in a
system.
We adopted ABMS so that we can model ECM
activities (Garcia, 2005) affected by various factors
such as consumer bahavior, competitive or
cooperative relationship (Lam, 2007) among
companies, and different adaptive strategies by
manufacturers. These factors are quite difficult to be
modelled using other conventional simulation
modeling techniques.
2.2 Methodology
2.2.1 The Basic Model
The basic model was built to address a question:
what is the impact on customers' satisfaction level
when limited resources are allocated differently
between necessary EC and initialized EC. Some
simple properties and rules are given to agents, and
we expect aggregate macro-scale behaviours or
trends emerging from the self-adaptive and
interactions between agents.
2.2.2 Hypothesis
We came up with six hypotheses concerning EC
factors, then built models to see whether these can
be supported or not under certain circumstance. In
future, we plan to use real data to see its validation.
ECM Hypotheses
H1. The effectiveness of a company's ECM is
positively correlated with the firm's market share.
H2. The degree of co-operation between
manufacturers and suppliers is positively correlated
with the performance of the firm.
H3. Changing the ratio between initialized ECs
between necessary ECs may lead to a different result
in a short period time.
H4. Initialized ECs are not important in gaining
market share.
H5. The level of EC frequency is positively
correlated with customers' satisfaction.
H6. ECM results are different for different types of
industry.
3 MODEL DESCRIPTION
3.1 Agents and Behaviour Rules
Agents represent autonomous decision-making
entities that interact with each other and/or with their
environment based on a set of rules. In a reasonable
environment, every agent would get its necessary
information, make adjustments on its behaviour
following the rule through iterative learning, and
pursue a certain goal or objective. Specific agents
used in our models are described next.
3.1.1 Manufacturer Agent
Manufacturers make similar products in a same
industry. Besides arranging daily regular production,
they receive ECRs driven by customers as well as
themselves. Then they evaluate and make decisions
on whether an EC is a necessary EC or an initialized
EC. They implement and track ECs. Also they
obtain feedback from the market to adjust their
strategy.
Different types of manufacturers have different
rules to govern their behaviour. In our models, there
is one control manufacturer who keeps its strategy
constant. It depends only on feedback of market
share and adjusts the ratio of resources used for
necessary ECs vs. Initialized ECs in order to
determine how to get maximum profits. Another
type of manufacturer uses a feed forward strategy to
act. The third kind of manufacturer is an intelligent
manufacturer that memorizes prior decisions and
results and learns to perform best.
3.1.2 Consumer Agent
Consumers may propose ECRs, and consumer
satisfaction is based on price, other consumers’
opinions, product quality and the level of continuous
improvement in product. These rational consumers
take best offers with highest personal satisfaction
(Kano, 1984). The interactions between different
kinds of agents and among similar kinds agents lead
to an aggregate macro scale behaviours or emerging
trends.
3.1.3 Supplier Agent
Suppliers keep in close touch with a manufacturer,
and many ECs may need supplier’s help to reduce
EC cost and lead time, or improve efficiency.
However, these will cost extra communication and
research expenses. A manufacturer may choose to
cooperate with other suppliers or not, considering
ICEIS 2009 - International Conference on Enterprise Information Systems
282

their potential profits as in the prisoners’ dilemma
scenario.
3.2 EC comparison
A summary of characteristics of two kinds of EC
(necessary EC and initialized EC) is given in Fig. 1.
Figure 1: Comparison of two kinds of ECs used in models.
3.3 Model Procedure
Figure 2 shows a general procedure used in the
models.
Figure 2: General model procedure.
First, we initialize the population of manufacturers
and consumers. Also, we determine the
manufacturers’ initial resource allocation value with
adaptiveness that limits the range of allocation
adjustment.
Every consumer has some probability to buy a
certain kind of product regardless of its vendor, and
the probability varies depending on people and
products. We also inject some influencing
consumers whose opinions have more weights on
other consumers.
An initial risk is determined based on different
industries’ attributes, but it can be changed during
simulation.
We take ECM tool quality and components
standardization into consideration in ECM attributes.
There exists one variable called "competition
degree" that means how fast potential manufacturers
may grow and enter into this market for competition,
and of course along with potential bankruptcy.
Output displays show the status of manufacturer
resource allocation of initialized ECs, market share
ratio comparison, profit comparison, number of
consumers who prefer initialized ECs. An agent map
shows consumer inclination switch by color change.
4 RESULTS
4.1 Preliminary Observations
The model results show that the initial ratio of
allocation plays a significant role. If one
manufacturer cannot obtain accurate information of
consumers’ preference or inclination unlike its
competitor, the performance in terms of market
share is very poor. Early bird catches the worm.
However, if other conditions are similar or the
same, a manufacturer a backward strategy may catch
up with others and even dominate the market. A feed
forward strategy works to some extent, but the
efficiency rely on initial difference between
forecasted data with real data. Intelligent agent
always performs best, since it obtains the advantage
of the pervious two strategies, makes distinctive
approach from learning experience. The strategy to
stay unchanged usually results in the worst
performance.
ECM tool and standardization play important
roles in EC process. A higher implementation level
of these can help to increase profits and market share
of the manufacturer. However, the model results
show that the significance of such ECM tool and
AGENT-BASED MODELING AND SIMULATION OF RESOURCE ALLOCATION IN ENGINEERING CHANGE
MANAGEMENT
283

standardization are not so great. Since an initial
strategy determines whether a manufacturer receives
the right information of trend from consumers, the
adaptability (Weiss, 1996) is the key for catching up
with a market leader and for keeping its market
share as well as its loyal consumers.
Other influencing factors such as advertisement,
consumers’ forum, etc., seem to make the
manufacturer easier to dominate the market if their
products attract most of the consumers in the
beginning. However, it does not mean that this
situation cannot be altered, even though the change
seems to be pretty hard. The difficulty for one
manufacturer to snatch the dominator position is
positively correlated with the value of influence-
probability.
Competition become much more fierce when
potential manufacturers' market shares grow at a
faster speed and are aggressive such as certain
suppliers who want to expand to become a
manufacturer. Eliminating manufacturers who
occupy very limited market share with low profit
through competition happens easily.
4.2 Conclusions on the Hypotheses
In this Section, the six hypotheses that we posed in
Section 2 are assessed based on the simulation
results.
H1. An effective ECM brings low cost and high
efficiency to a company, which leads to higher
profits and market share.
H2. Our model focused on the relationship
between supplier & manufacturer using the game
theory of prisoner’s dilemma to see the consequence.
A higher level of cooperation helps both of them,
but especially the manufacturer.
H3. The ratio between initialized ECs and
necessary ECs tells what a manufacturer emphasizes.
However, the impact is observable only after some
time.
H4. Initiated EC may or may not be a major
contributor toward gaining market share, depending
on the circumstances such lead time. If a lead time
for EC implementation is relatively short, the
initialized EC makes little impact. However, if
competition is intense enough, the pressure from the
market and competitors may force manufacturer to
make initialized ECs more frequently, thus giving
consumers more satisfaction.
H5. Up to a certain point, increasing EC
frequency helps to attract more loyal consumers.
However, very frequent ECs introduce more
disruptions to the manufacturing system leading to
worse performance in market share gain.
H6. We considered EC risk, EC lead time and
EC complexity to differentiate different industries. A
high risk causes two extreme phenomena.
Manufacturers may gamble to pursue profits even
though they are sometimes temporary profit. Or they
may stay with a conservative strategy to keep
foreseeable market share. While at low EC risk, the
competition is being encouraged.
5 CONCLUSIONS
The model results confirm that it is useful to classify
ECs into initialized ECs and necessary ECs. Also,
the competitive nature of a market influences how a
firm should emphasize necessary EC vs. initialized
EC. The greater the competitions are, the greater the
need to emphasize initialized ECs exists. The
situation will necesseciate the adoption of new
technologies that promote customer satisfaction to
excitement rather than just satisfaction.
Another interesting result is that intelligent
manufacturer who combines forecasting and
feedback strategy and learns from past experience
performs best in most cases. Still, adaptive feed
forward and feedback strategy works even better in
some cases. A possible explanation for this
phenomenon is that even though intelligent
manufacturer learns from past, some changes happen
without any foreseeable notices. As a result, the
adaptive ones outplay intelligent ones since they
simply rely on difference between expectation and
reality.
REFERENCES
Garcia R., 2005. Uses of Agent-Based Modeling in
Innovation/New Product Development. Journal of
Product Innovation Management, 2005, Vol 22, 380-
398.
Macal C., North M., 2005. Tutorial on agent-based
modeling and simulation, Proceedings of the 2005
Winter Simulation Conference, 2005, 2-15.
Moon, Y.B., 2007. Enterprise Resource Planning (ERP): a
review of the literature, International Journal of
Management and Enterprise Development, Vol. 4, No.
3, 2007, pp. 235-264.
Nadia B., Gregory G., Vince T., 2006.Engineering change
request management in a new product development
process, European Journal of Innovation Management
Vol. 9 No. 1, 2006, 5-19.
ICEIS 2009 - International Conference on Enterprise Information Systems
284