Crash failure detection in asynchronous agent communication languages

摘要

Agent Communication Languages (ACLs) have been developed to provide a way for agents to communicate with each other supporting cooperation in Multi-Agent Systems (MAS). In the past few years many ACLs have been proposed for MAS and new standards are emerging such as the ACL developed by the Foundation for Intelligent Physical Agents (FIPA). Despite these efforts, an important issue in the research on ACLs is still open and concerns how these languages should deal with failures of agents in asynchronous MAS. The Fault Tolerant Agent Communication Language (\(\mathbb{FT}\)-\(\mathbb{ACL}\)) presented in this paper addresses this issue dealing with crash failures of agents. \(\mathbb{FT}\)-\(\mathbb{ACL}\) provides high-level communication primitives which support a fault-tolerant anonymous interaction protocol designed for open MAS. We present a formal semantics for \(\mathbb{FT}\)-\(\mathbb{ACL}\) and a formal specification of the underlying agent architecture. This formal framework allows us to prove that the ACL satisfies a set of well defined knowledge-level programming requirements. To illustrate the language features we show how \(\mathbb{FT}\)-\(\mathbb{ACL}\) can be effectively used to write high-level executable specifications of fault tolerant protocols, such as the Contract Net one.