{"references": ["K.M. Chandy and J. Misra. How processes learn. Distributed Computing\n1(1), pp. 40-52. Springer-Verlag, 1986.", "M.R. Genesereth and N.J. Nilsson. Logical Foundations of Artificial\nIntelligence. Morgan Kaufmann, Palo Alto, CA, 1987.", "M. Hennesy and R. Milner. Algebraic laws for nondeterminism and\nconcurrency. Journal of the ACM 32(1), pp. 137-161. 1985.", "L. Lamport. A simple approach to specifying concurrent systems.\nCommunications of the ACM 32(1), pp. 32-45. 1989.", "B. van Linder,W. van der Hoek, and J.J. Ch. Meyer. Formalising abilities\nand opportunities of agents. Fundamenta Informatica, 34, (1,2), pp. 53-\n101, 1998.", "N. Lynch and M. Tuttle. An introduction to input/output automata. CWI\nQuaterly 2(3), pp. 219-246. 1989.", "G. Ndjatou. Minimizing agent specifications using a logic of knowledge\nand actions. Journal of Logic and Computation , vol. 11, No. 2, pp.\n337-354, Oxford University Press, 2001.", "G. Ndjatou. Modelling Objects, Knowledge and Learning in Distributed\nObject-Based Systems. Ph.D thesis, Dept. of Computer Science, CUNY\nGraduate School, New York, NY, Feb. 93. Also appeared as Technical\nReport N0 TR-93-04-02, CUNY Graduate School, New York, NY, April\n93.", "G. Ndjatou. Modelling objects and distributed object-based systems. In\nproceedings of the 10th Israeli Symposium on AI and Computer Vision ,\npp. 39-49, Ramat Gan, Israel, December 1993.\n[10] N. Reingold, D.W. Wang and L.D. Zuck. Games I/O automata play. Concur\n-92, Lecture Notes in Computer Science, vol. 630, W.R Cleaveland,\ned. pp. 325-339. Springer-verlag, 1992.\n[11] G. Weiss, ed., Multiagent Systems, A Modern Approach to Distributed\nArtificial Intelligence. The MIT Press, 1999."]}

We propose a formal framework for the specification of the behavior of a system of agents, as well as those of the constituting agents. This framework allows us to model each agent-s effectoric capability including its interactions with the other agents. We also provide an algorithm based on Milner-s "observation equivalence" to derive an agent-s perception of its task domain situations from its effectoric capability, and use "system computations" to model the coordinated efforts of the agents in the system . Formal definitions of the concept of "behavior equivalence" of two agents and that of system computations equivalence for an agent are also provided.