J. Software Engineering & Applications, 2010, 3: 404-408
doi:10.4236/jsea.2010.34045 Published Online April 2010 (http://www.SciRP.org/journal/jsea)
Copyright © 2010 SciRes JSEA
Intelligent Supply Chain Management
Mohammad Zubair Khan1, Omar Al-Mushayt1, Jahangir Alam2, Jorair Ahmad1
1Faculty of Computer Science and Information System, University of Jizan, Jizan, Kingdom of Saudi Arabia; 2University Women’s
Polytechnic, Faculty of Engineering and Technology, Aligarh Muslim University, Aligarh, India.
Email: Zubair.762001@gmail.com, oalmushayt@yahoo.com, Jahangir.uk786@yahoo.co.uk
Received December 27th, 2009; revised January 31st, 2010; accepted February 5th, 2010.
Fuzzy Logic is used to derive the optimal inventory policies in the Supply Chain (SC) numbers. We examine the per-
formance of the optimal inven tory policies by cutting the costs and increa sing the supply chain management efficiency.
The proposed inventory policy uses multi-agent and Fu zzy logic, and provides managerial insights on the impact of the
decision making in all the SC numbers. In particular, we focus on the way in which our agent purchases components
using a mixed procurement strategy (combining long and short term planning) and how it sets its prices according to
the prevailing market conditions and its own inventory level (because this adaptivity and flexibility are key to its suc-
cess). In modern global market, one of the most important issues of the supply chain (SC) management is to satisfy
changing customer demands and enterprises should enhance the long-term advantage through the optimal inventory
control. In this paper an intelligent multi-agent system to simulate supply chain man agement has bee n dev elo ped .
Keywords: SCM Supply Chain Management, Customer Agent, RFQ, Component Agent
1. Background and Introduction
1.1 Supply Chain Management
Supply chains encompass the companies and the business
activities needed to design, make, deliver, and use a
product or service. Businesses depend on their supply
chains to provide them with what they need to survive
and thrive [1-3]. Every business fits into one or more
supply chains and has a role to play in each of them [3,4].
A multiechelon supply chain is illu strated in Figure 1. It
includes different types of flows i.e. financial flow, In-
formation flow and material flow. Third party logistics
(3PL) services providers (3PL is an organization that
manages and executes a particular logistics function,
using its own assets and resources, on behalf of another
company.) handle the inbound and outbound logistics for
the shipper (The person or company who is usually the
supplier or owner of commodities shipped). The inbound
logistics take care of the material management while the
outbound logistics deal with physical distribution of final
products. The term “supply chain management” arose in
the late 1980s and came into widespread use in the 1990s.
Prior to that time, businesses used terms such as “logis-
tics” and “operations management” instead. Some defini-
tions of a supply chain are offered below:
“A supply chain is the alignment of firms that bring
products or services to market.”—Lambert, Stock, and
Ellram (1998).
“A supply chain consists of all stages involved, di-
rectly or indirectly, in fulfilling a customer request. The
supply chain not only includes the manufacturer and
suppliers, but also transpo rters, warehouses, retailers, and
customers themselves.”—Chopra and Meindl (2001).
“A supply chain is a network of facilities and distri-
bution options that performs the functions of procure-
ment of materials, transformation of these materials into
intermediate and finished products, and the distribution
of these finished products to customers.”—Ganeshan and
Harrison (1995).
Master strategist and a skillful general Napoleon once
remarked, “An army marches on its stomach.” Which
holds true in business, as it moves on physical flow of
materials, another saying that goes is also valid for the
business, “Amateurs talk strategy and professionals talk
logistics” [5-7].
Thus, we can say that Supply chain management is a
set of approaches used to efficiently integrate suppliers,
manufacturers, warehouses, and customers so that mer-
chandise is produced and distributed at the right quanti-
ties, to the right locations, and at the right time in order
to minimize system wise costs while satisfying service-
level requirements.
Supply chain management (SCM) is the task that
moves in a process from supplier to manufacturer to
wholesaler to retailer to consumer [1,4]. Supply chain
management involves coordinating and integrating these
Intelligent Supply Chain Management 405
Inbound LogisticsO utbound Logistics
Material Mana gementPhysical Distribution
3PL Services Providers
Flow of Information
Flow of Goods
Flow ofFinancealon g with Information
Figure 1. The supply chain process
flows both within and among companies [2,8]. It is said
that the ultimate goal of any effective supply chain man-
agement system is to reduce inventory (with the assump-
tion that products are available when needed).
Supply chain management flows can be divided into
three main flows:
The product flow
The information flow
The finances fl ow
The need of Supply chain Management in today’s
scenario is:
Millions of dollars at stake!
Excess Inventory costs
Excess freight charges
Lost sales/Stock outages
Wasted time and energy
Extra staff
Customer dissat isfaction – p rivatization
Capital costs
Real Estate Costs
Static and unresponsive SC policies
Large inventories
Unreliable deliveries
1.2 Fuzzy Logic
Fuzzy logic is a form of multi-valued logic derived from
fuzzy set theory to deal with reasoning that is approxi-
mate rather than precise. Just as in fuzzy set theory the
set membership values can range (inclusively) between
any values, in fuzzy logic the degree of truth of a
statement can range between any values and is not
constrained to the two truth values {true, false} as in
classic predicate logic [1]. And when linguistic variables
are used, these degrees may be managed by specific
functio n s, a s discussed be l ow.
Fuzzy Set Theory defines Fuzzy Operators on Fuzzy
Sets. The problem in applying this is that the appropriate
Fuzzy Operator may not be known. For this reason,
Fuzzy logic usually uses IF-THEN rules, or constructs
that are equivalent, such as fuzzy associative matrices.
Rules are usually expressed in the form: IF variable IS
property THEN action.
For example, an extremely simple temperature regulator
that uses a fan might look like this: If temperature IS
very cold THEN stop fan. If temperature is cold THEN
turn down fan. If temperature IS normal THEN maintain
level. IF temperature IS hot THEN speed up fan. Notice
there is no “ELSE”. All of the rules are evaluated,
because the temperature might be “cold” and “normal” at
the same time to different degrees. This paper proposes a
multi-agent system to simulate supply chain management
using the concepts of fuzzy logic. The rest of the paper is
organizes as follows—Section 2 presents the proposed
model, Methodology adopted conducting the experi-
ments and results so obtained have been discussed in
Section 3. Section 4 concludes the paper.
2. Proposed Model
The proposed SCM model has been depicted in Figure 2.
In our model we are introducing the customer agent and
Customer RFQ (Request for Quantities). The customer
agent receives customer RFQ requesting a quantity of a
particular commodity for delivery on a specified day.
The customer agent is the key component while compo-
nent agent is responsible for dealing with the component
suppliers and aims to ensure that there are always suffi-
cient components in stock to address the customers’
changing demand for finished products.
In our model we propose customer agent only and
leave the component agent for future studies.
2.1 The Customer Agent
The customer agent is the key component in our model
(because we believe that offering the appropriate price at
the right time is vital for success). If the price is too low,
the agent will receive a low profit and if it is too high it
will fail to win any orders (because the natural tendency
of customers is to always choose the lowest offer price
among those they receive). Given this, the key challenges
are to determine which customer RFQ to bid for and at
Copyright © 2010 SciRes JSEA
Intelligent Supply Chain Management
Figure 2. The SCM model
what price. To achieve this, we use fuzzy reasoning [9,10]
to determine how to set prices according to the agent’s
inventory level, the market demand and the time into
2.2 Choosing Customer RFQ’s
The customer agent receives customer RFQs requesting a
quantity of a particular commodity for delivery on a
specified day [2,8,11]. When selecting which RFQs to
respond to, Our ISCM rates them according to the poten-
tial profit that they may bring and according to the in-
ventory it holds. The latter inventory driven strategy of-
fers customers commodity according to what is currently
available and also what can be produced given the deliv-
ery date of pending components orders. In more detail,
suppose a customer RFQ is represented as a tuple (i, q,
pres, cpenalty, pbase), where i is the type of commodity
the customer wants, q > 0 is the quantity, pres > 0 is the
reservation price (maximum it will pay), cpenalty > 0 is
the fine paid if the computers are not delivered on time,
and ddue is the desired delivery date. On each day, the
customer agent receives a bundle of such RFQs and sorts
them in the decreasing order of the profit margin of the
type of commodity requested. Here pbase is the cost of
buying components (sum of the buying price for each
component). The intuition is that the agent will first serve
customers with high profit margins and low penalties.
This is because the higher the pres, the more profit will
be made (compared to selling the same product to a cus-
tomer with a low pres). At the same time, the agent also
wants to avoid getting high penalty orders because of the
inherent uncertainties that exist in the scenario.
Profit Margin = Pres-Pbase-(cpenalty/q)
3. Methodology
We are applying the Fuzzy logic if-else algorithm using
the following rules and simulated it using available data.
Here we define fuzzy rules and then revisited the rules
for our g oa l.
3.1 Fuzzy Rules for Calculating Offer Prices
R1: if D is high and I is low then r1 is very-big
R2: if D is medium and I is high then r2 is
R3: if D is low and I is high then r3 is small
3.2 Fuzzy Rules Revisited
If D is high I is High E is far then r1` is big.
If D is high I is High E is in between then r2` is big.
If D is high I is High E is near then r3` is big.
3.3 The Component Agent
The component agent is responsible for dealing with the
component suppliers and aims to ensure that there are
always sufficient components in stock to address the cus-
tomers’ changing demand for finished products. In doing
so, it addresses a challenge that is common to all supply
chains facing dynamically changing customer demand.
That is, it must procure components at a low cost, whilst
simultaneously maintaining a minimal component inven-
tory in order to reduce the daily storage cost and also the
possibility of being left with redundant stock if customer
demands changes.
Types of procu re me nts
Far future proc urement
Near Future Procurement
3.3.1 Far Future Procurement
Of the two strategies, the far future procurement one is
the simpler. For this, the agent assumes that in the far
future, there will be a daily minimum need, and thus
checks whether there is sufficient current and pending
Copyright © 2010 SciRes JSEA
Intelligent Supply Chain Management407
component inventory to meet this need. If not, it submits
an RFQ for a fixed amount to the relevant supplier, re-
questing delivery on the date at which the predicted in-
ventory falls below the daily minimum need.
3.3.2 Near Future Procurement
The near future procurement strategy is more complex. It
consists of two elements, a daily demand predictor that
predicts the future demand of components, and a market
tracker that generates the RFQs to be sent to the suppliers,
both to order actual components required and to test the
market to discern the most profitable order lead time and
set appropriate reserve price
3.4 Demand Predictor
As described above, the agent buys components for the
near future based on a prediction of customer demand.
Now, according to the game specification, the number of
RFQs that each agent receives from the customers is de-
scribed by three independent random walks; one for each
market segment (finished products are classified into
three such segments: high, mid and low range). In more
detail, the number of RFQs that an agent receives, within
a single market segment, on day d is denoted by Nd, and
is drawn from a Poisson distribution whose expected
value is given by the parameter, Qd. Thus, for each mar-
ket segment, Nd = Poisson(Qd)
Having predicted the number of RFQs that will be re-
ceived, within each market segment, on each day within
the near future, the agent then calculates the expected
daily usage of each component type (Did).
3.5 Price Tracker
The price tracker acts to maintain an estimate of the cur-
rent market price of the components. Due to the behav-
iors of the competing agents, this market price depends
on the due date with which components are requested.
For example, if the competing agents are ordering com-
ponents with very short lead times, then the supplier will
have little spare capacity, and thus, the corresponding
offer prices that the agent receives will be greater than
those of orders wit h l ong l ead t im e s.
3.6 Factory Agent
One of the main challenges for the factory agent is
scheduling what to produce and when to produce it (i.e.,
how to allocate supply resources and factory time).
This strategy involves manufacturing commodities
according to customer orders and satisfying orders with
an earlier delivery date. Now, since the computers stored
in the factory will be charged storage cost, each order
will be delivered as soon as it is filled. The agent builds
the commodity according to the customers’ orders it has
obtained (which has the advantage of ensuring that the
factory always produces the needed computers on time).
However, if on any day, there are still free factory as-
sembling cycles available, and the numbers of finished
PCs in stock are below a certain thresho ld, then the agen t
produces additional PCs of each kind uniformly (subject
to the availability of components) in order to maximize
the factory utilization. It is critical that this threshold is
set appropriately; a high threshold will lead to excessive
finished PC inventory, which may be hard to sell if de-
mand is low.
3.7 Simulated Results
Margin Inventory End of sea-
son Predictors
High High Far yes
High High In Between yes
High High Near yes
High Medium Far yes
High Medium In Between yes
High Medium Near no
High Low Far yes
High Low In Between no
High Low Near no
Medium High Far yes
Medium High In Between yes
Medium High Near yes
Medium Medium Far yes
Medium Medium In Between yes
Medium Medium Near no
Medium Low Far no
Medium Low In Between no
Medium Low Near no
Low High Far no
Low High In Between no
Low High Near yes
Low Medium Far no
Low Medium In Between no
Low Medium Near no
Low Low Far no
Low Low In Between no
Low Low Near no
4. Conclusions
This mixture of baseline and opportunistic purchasing
behavior is a common strategy in this domain and the
technology we develop for achieving this can be readily
transferred. Second, we believe our pricing model tech-
nology will also be useful in real SCM applications
where just undercutting competitors’ prices can signifi-
cantly improve profitability. Specifically, to apply our
model in other domains, the designers of the rule base
would need to adapt the fuzzy rules to reflect the factors
that are most relevant. Now we believe that customer
demand and inventory level are highly likely to be criti-
cal factors for almost all cases and thus these rules can
remain unaltered.
By using different rule bases, different factors can eas-
ily be incorporated (as we did here, in order to handle the
additional need to reduce inventory towards the end of
the season). The purpose model also will use in this area
Copyright © 2010 SciRes JSEA
Intelligent Supply Chain Management
Copyright © 2010 SciRes JSEA
focuses on the component agent. We would like to im-
prove it so that it can adap t the quantity for far fu ture and
near future procurement automatically between the sea-
sons according to the procurement behaviors employed
by the opponents.
[1] J. Collins, R. Arunachalam, et al., “The Supply Chain
Management Game for the 2005 Trading Agent Competi-
tion,” Technical Report CMU-ISRI-04-139, School of
Computer Science, Carnegie Mellon University, Pitts-
burgh, December 2004.
[2] S. D. Levi, P. Kaminsky and S. E. Levi, “Designing and
Managing the Supply Chain,” McGraw-Hill, Illinois,
[3] D. Pardoe and P. Stone, “TacTex-03: A supply Chain
Management Agent,” SIGecom Exchanges: Special Issue
on Trading Agent Design and Analysis, Vol. 4, No. 3,
2004, pp. 19-28.
[4] K. Kumar, “Technology for Supporting Supply-Chain
Management,” Communications of the ACM, Vol. 44, No.
6, pp. 58-61, 2001.
[5] D. Pardoe and P. Stone, “Predictive Planning for Supply
Chain Management,” Proceedings of International Con-
ference on Automated Planning and Scheduling, to appear.
[6] M. Sugeno, “An Introductory Survey of Fuzzy Control,”
Information Sciences, Vol. 36, 1985, pp. 59-83.
[7] M. Wellman, J. Estelle, S. Singh, et al., “Strategic Inter-
actions in a Supply Chain Game,” Computational Intelli-
gence, Vol. 21, No. 1, 2005, pp. 1-26.
[8] M. He, H. F. Leung and N. R. Jennings, “An ARTMAP
Based Bidding Strategy for Autonomous Agents in Con-
tinuous Double Auctions,” IEEE Transactions on
Knowledge and Data Engineering, Vol. 15, No. 6, 2003,
pp. 1345-1363.
[9] R. Arunachalam and N. Sadeh, “The Supply Chain Trad-
ing Agent Competition,” Electronic Commerce Research
and Applications, Vol. 4, No. 1, 2005, pp. 63-81.
[10] J. Collins, R. Arunachalam, N. Sadeh, J. Ericsson, N.
Finne and S. Janson, “The Supply Chain Management
Trading Agent Competition,” Technical Report CMU-
ISRI-04-139, Carnegie Mellon University, Pittsburgh,
[11] M. He, N. R. Jennings and H. Leung, “On Agent-Medi-
ated Electronic Commerce,” IEEE Transactions on
Knowledge and Data Engineering, Vol. 15, No. 4, 2003,
pp. 985-1003.