Talks » Class » MATES09OBS

Exploiting Agent Observability in Multi-Agent System Programming

Alessandro Ricci, Michele Piunti
Along with direct interaction models, relying on message-based explicit communication, indirect interaction models can play an important role for designing coordination strategies in Multi-Agent Systems (MAS). Among indirect models, those based on agent observability such as over-sensing achieve coordination by letting agents observe and possibly reason about (each) others' actions and state. In this paper we describe how over-sensing and, more generally, agent observability mechanism can be used in practice to program MAS in the context of artifact-based environments by exploiting the support provided by CArtAgO technology.
Presented by Michele Piunti
German Conference on Multi-Agent Systems (MATES 2009)
09/09/2009