Evaluation of Environment Contextual Services in Multiagent Systems

The environment is a powerful first-order abstraction in Multi-AgentSystems (MAS), as well as a critical building block. The agents interact in theirenvironment and the effects of their actions are observed and evaluated throughthis environment. The local complexity of the agents depends on its managementof the interaction and action processes. If the environment carries out a part ofthis processes, the complexity of the agents is reduced. This delegating processimplies a centralization of a part of the MAS computations inside the environmentand therefore a flexible way to exchange information and to coordinate the agents.In this paper, we present the modeling of an environment which supports bothcommunication services and simulation services : multi-party communications(communication) and contextual activation (simulation). We evaluate the cost ofthese environment services and compare it to the execution of the same tasksinside the agents. The evaluation and comparison are done theoretically and em-pirically for communication and simulation. We also investigate the clustering ofthe agents in several environments.ouinonouirechercheInternationa