Agent Communication Frameworks and Verification

This paper develops a general agent communication framework which allows us to define several different notions of verification and to investigate if an agent communication language is verifiable. The framework is sufficiently general to accommodate communication languages based on agents’ mental states as well as those based on social states of the multi-agent system. For this purpose an existing computational model is employed to represent agent states and extended to represent social states. We use this framework to identify the types of languages that are verifiable in open systems where agents’ internals are kept private.