Agents 'Interaction Capacity' Language (AICL)

摘要

In this paper we present a multi-agent system design methodology and an agents Interaction Language. The methodology covers the next points: The agent architecture for defining the agent characteristics and functioning, a formal model to specify agents, and the object language to describes agents' interaction. The agent architecture is concordant with classical architecture called deliberative architecture, which contain an explicit symbolic model of the world and decisions arc taked through logical reasoning. That is, our architecture may be regarded as a subset of BDI architecture, which are included into ICE Architecture as a special case, As a forma] method for specifying agent's interaction process, we propose using an Agent Meta Language (AML). Our AML is an Agent Logic (AL) supported by a belief temporal logic. AICL is a language for developing Multi-Agent applications that is, cooperative applications. AICL's sentences are constructed using KQML primitive messages, in this way; language can be conseived as a layer upon KQML. AICL's sentences are used by agents for interchanging messages containing information useful to converge towards the solution of the problem entrusted to the system. Object language AOL (Agent Object Language) generated by AICL uses the paradigm "Interaction Capacity".