Using semantic causality graphs to validate MAS models

Abstract

Multi-agent systems methodologies define tasks and tools covering all the development cycle. Modelling and implementation activities are well supported and experimented. However, there has been few effort invested in debugging and validation techniques for MAS compared to modelling and implementation stages. This paper addresses the lack of tools to debug and validate a MAS. We propose a visual tool to debug and validate some multi-agent system specification, capturing causality among messages sent by agents and showing semantic information related to interaction protocols used by agents. © 2007 Springer-Verlag Berlin Heidelberg.