Title: Sketching a methodology for efficient Supply Chain Management agents enhanced through Data Mining

Authors: Andreas L. Symeonidis, Vivia Nikolaidou, Pericles A. Mitkas

Addresses: Electrical and Computer Engineering Department, Aristotle University of Thessaloniki, 54 124 Thessaloniki, Greece. ' Electrical and Computer Engineering Department, Aristotle University of Thessaloniki, 54 124 Thessaloniki, Greece. ' Electrical and Computer Engineering Department, Aristotle University of Thessaloniki, 54 124 Thessaloniki, Greece

Abstract: Supply Chain Management (SCM) environments demand intelligent solutions, which can perceive variations and achieve maximum revenue. This highlights the importance of a commonly accepted design methodology, since most current implementations are application-specific. In this work, we present a methodology for building an intelligent trading agent and evaluating its performance at the Trading Agent Competition (TAC) SCM game. We justify architectural choices made, ranging from the adoption of specific Data Mining (DM) techniques, to the selection of the appropriate metrics for agent performance evaluation. Results indicate that our agent has proven capable of providing advanced SCM solutions in demanding SCM environments.

Keywords: data mining; intelligent agents; supply chain management; SCM; agent-oriented methodology; bidding; forecasting; agent-based systems; multi-agent systems; MAS; trading agent competition; performance evaluation.

DOI: 10.1504/IJIIDS.2008.017244

International Journal of Intelligent Information and Database Systems, 2008 Vol.2 No.1, pp.49 - 68

Published online: 20 Feb 2008 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article