Agent-based computing in Arti cial Intelligence has given rise to a number of diverse and competing proposals for agent programming languages. Agents, in the sense we are using it, are complex mental entities consisting of beliefs, goals, and intentions. For several reasons it has been di cult to evaluate and compare the di erent proposals for agent languages. One of the main reasons, in our opinion, is the lack of a general semantic framework which provides a suitable basis for language comparison. Our aim is to make as much use as possible of formal methods from the area of programming semantics. In this paper, we give a formal embedding of the agent language AgentSpeakL in our own agent language 3APL. To this end we de ne a notion of simulation based on the formal operational semantics of the languages. The main result of the paper is a proof that 3APL can simulate AgentSpeakL. As a consequence, 3APL has at least the same expressive power as AgentSpeakL. The comparison yields some ...
Koen V. Hindriks, Frank S. de Boer, Wiebe van der