Information Technology Reference
In-Depth Information
agent to adapt its price setting to the prevailing market situation, its own internal state
(inventory level) and the time that has elapsed. At their core, these techniques employ
fuzzy reasoning in order to allow the agent to adapt its prices daily so that it can fully
exploit its production capacity, while still maximising its revenue by selling at appro-
priate prices. Previously, fuzzy techniques have been successfully applied to solve the
problems of automated auction [3,5], TAC classic (SouthamptonTAC [3]) and negotia-
tion [7]. We believe fuzzy logic provides an effective tool to cope with the uncertainty
inherent in a complex decision making problem (e.g. the supply chain context) and to
make trade-offs between the variants of attributes (e.g. price and quantity). Also, fuzzy
rules are the most visible and interpretable manifestation of this approach and have been
successfully used in a variety of areas [10].
The remainder of the paper is organized as follows. Section 2 outlines the TAC SCM.
Section 3 presents our agent. Section 4 evaluates the performance of the agent. Finally,
Section 5 concludes.
2TheTACSCMGame
In this game, six agents (competition entrants) compete with one another to procure
raw components and fulfil customer orders for assembled PCs. Each PC is assem-
bled from four components: CPU, motherboard, memory and hard disk (e.g. a PC
with a 2GHz IMD processor with 1GB memory and a 300GB hard drive). Each agent
is able to produce any of the 16 distinct computer types (different PC types require
a different number of production cycles) and is limited to a capacity of 2000 cycles
daily.
The agents operate simultaneously in separate markets to buy components from a
number of suppliers and to sell PCs to customers. Both of these markets operate as
follows: (i) the buyer issues Request For Quotes (RFQs) to the sellers; (ii) the sellers
respond to the RFQs with offers detailing the price, quantity or delivery date; and (iii)
the buyer sends orders to accept offers.
Consequently, on each of the 220 simulation days of the game, agents receive from
the customers a new set of RFQs and, in response to previously sent offers, they receive
orders for assembled computers. Likewise, component suppliers that were previously
sent RFQs respond with offers. Thus, in each day of the game (lasting 15 seconds), the
agent must decide on the following: (i) which new supplier RFQs to issue and which
supplier offers to accept; (ii) which customer RFQs to respond to, and what price to
offer; and (iii) how to schedule the production and delivery of PCs given the availability
of components, the limited capacity of the factory and the delivery deadlines of pending
orders.
An agent spends money on buying the components, paying for the storage of both
components and PCs, paying penalties if it defaults on a promised delivery date and
paying overdraft penalties if it is in debt to the bank. The agent earns money by selling
PCs and receives interest from the bank if its balance is positive. Success of an agent is
measured in terms of its profit (
i.e.
, its bank balance at the end of the game).
