On Formal Modeling of Agent Computations

This abstract describes a comparative study of three formal methods for modeling and validating agent computations. The experiment is part of a joint project by researchers in MIT's Theory of Distributed Systems research group and NTT's Cooperative Computing research group. Our goal is to establish a mathematical and linguistic foundation for describing and reasoning about agent-style distributed systems. Agents are autonomous software entities that cooperate with other agents in carrying out delegated tasks. Key features of agent systems include: (1) They are dynamic, in that they allow run-time creation and destruction of processes, run-time modication of communication capabilities, and mobility. (2) The state of an agent typically includes a knowledge base, which keeps track of facts that it \knows" (knowledge set), and facts that it \believes" (belief set). However, we do not emphasize this latter structure in this paper. Our experiment involves examining the power of three formal methods|I/O automata, knowledgebased programs, and Nepi (a variant of the calculus)|in studying di erent agent applications. Here, we summarize our experience modeling a basic problem arising in e-commerce systems, using the three formalisms.