An Agent-Based Solution for Dynamic Supply Chain Management

Vedran Podobnik (University of Zagreb, Croatia)

Ana Petric (University of Zagreb, Croatia)

Gordan Jezic (University of Zagreb, Croatia)

Abstract: Supply chain management (SCM) deals with planning and coordinating activities such as material procurement, product assembly, and the distribution of manufactured products. This paper offers an agent-based solution as a potentially adequate approach for the automation of supply chain management. The greatest obstacle in SCM research is obtaining benchmark designed solutions since it is difficult to simulate real business environments, while live testing in real-world systems is not an option. The Trading Agent Competition Supply Chain Management (TAC SCM) scenario provides a unique testbed for studying and prototyping SCM agents by providing a challenging game environment where competing agents engage in complex decision-making activities with the purpose of maximizing their profit. In this paper, we describe the TAC SCM environment and present the main features of the CrocodileAgent, our TAC SCM 2007 entry. Additionally, the CrocodileAgent’s performance in the competition, as well as in a series of controlled experiments, is discussed.

Keywords: electronic markets, multi-agent simulation, software agents, supply chain management, trading agents

Categories: I.2.1, I.6.0, J.7, K.4.4