Information Technology Reference
In-Depth Information
An Analysis of the 2004 Supply Chain Management
Trading Agent Competition
Christopher Kiekintveld, Yevgeniy Vorobeychik, and Michael P. Wellman
University of Michigan
Artificial Intelligence Laboratory
Ann Arbor, MI 48109-2110 USA
ckiekint@umich.edu
Abstract. We present and analyze results from the 2004 Trading Agent Com-
petition supply chain management scenario. We identify behavioral differences
between the agents that contributed to their performance in the competition. In
the market for components, strategic early procurement remained an important
factor despite rule changes from the previous year. We present a new experimen-
tal analysis of the impact of the rule changes on incentives for early procurement.
In the finals, a novel strategy designed to block other agent's access to suppliers at
the start of the game was pivotal. Some agents did not respond effectively to this
strategy and were badly hurt by their inability to get crucial components. Among
the top three agents, average selling prices in the market for finished goods were
the decisive difference. Our analysis shows that supply and demand were key fac-
tors in determining overall market prices, and that some agents were more adept
than others at exploiting advantageous market conditions.
1
Introduction
The Trading Agent Competition (TAC) provides an international forum for researching
the design and analysis of automated trading agents. A new scenario in supply chain
management (TAC/SCM) debuted in 2003 [1]. We will not describe the scenario here,
but direct the reader to the game specification for details [2]. Studying the outcomes of
these competitions is a valuable exercise that helps us to better understand the strengths
and weaknesses of current approaches. Here we present and analyze results from the
final round of the 2004 tournament. Our primary objective is to determine the impor-
tant behavioral factors that distinguished the agents and contributed to their relative
performance.
We start by presenting the main results from the 2004 finals. In Section 3 we consider
procurement strategies. We discuss the role of strategic early procurement, and present
new experimental analysis of the effect of the rule changes on early procurement. A
novel blocking strategy was used in the final round, and we discuss the problems this
cause for some of the agents. In Section 4 we consider the PC sales market. Agents had
widely varying average selling prices (ASPs) for PCs, and this was a deciding factor
between the top three agents. Further analysis identifies four factors strongly correlated
with market ASPs, and reveals that the agents differed in their ability to target profitable
markets. We conclude with a summary of the strengths and weaknesses of each agent.
Search WWH ::




Custom Search