Comparing and Evaluating Organizational Models: A Multi-agent Programming Contest Case Study

Abstract

An important subset of multi-agent systems MAS are based on organizational models. These models try to define pre-defined intended agent interaction patterns. Given an application domain, however, the choice of a particular organizational model that better solves the problem is still an open problem. In order to guide this choice, a MAS developer must have the opportunity to test distinct organizational models easily. In this work, we compare and evaluate different organization models of a MAS, whose goal is to evolve in the "Agents on Mars" scenario proposed in the Multi-Agent Programming Contest MAPC.