Information Technology Reference
In-Depth Information
Figure 4. Agent trading diagram
P
P
t
t
t
t
P
P
t
P0
t
0
t
P
P
t
t
t
P
P
It is a tradition that DP formulations are treated as elegant descriptions of the decision problem,
which is then solved by
heuristic
methods. A
heuristic
method is a reasonable and logical method of
solving the problem in a manner similar to the mathematical approach, but not necessarily using the
mathematical approach, or even producing the optimal solution as given by the mathematical approach.
Described below are heuristic solution approaches as a proposed
agent based transaction system
(ABTS)
to the trading problem.
The
agent based transaction system
(ABTS) goes beyond the basic DP decision process by using
cooperative actions within the community to maximize transaction volume and utility for both buyer
and seller agents. Actions within the community, such as message passing and cooperation between
agents in matching an offer in buying or selling to maximize the volume, are not specified in the DP
formulation. In this approach, heuristic algorithms are described that include message passing between
agents to match an offer by using part quantities from cooperating agents.
In the real world of supply-chain transactions, split quantities represent higher costs. A seller agent
stuck with a small remaining quantity may be stuck with it for a long time with no buyer available. It
means increasing administrative costs and increasing inventory costs, until a buyer agent is found. There
is also the uncertainty that a buyer agent will be found, and the remaining inventory will not go into
salvage. For a buyer agent, a split quantity implies a small quantity to be purchased at a higher price,
if a seller agent can be found to sell such a small quantity. This work does not consider negotiations
of prices and costs. Instead, the issues of match and mismatch of quantities have been made issues of
Search WWH ::
Custom Search