Abstract：Co-evolution is a kind of good method that can solve multi-agent collaboration after symbolism and behaviorism method. In this paper, we put forward the multi-agent architecture and the coordination mechanism based on co-evolution and establish the multi-agent system simulation framework. Then, we design the class library for the simulation framework using object oriented method. At last, we make an experiment on the "predator-prey" problem.
[1] Nicholas R. Jennings, Katia Sycara, Michael Wooldridge. A Roadmap of Agent Research and Development. In: Autonomous Agents and Multi-Agent Systems,1998, 1,7,38 [2] Burt Wilsker. A Study of Multi-Agent Collaboration Theories. Information Science Institue research report#ISI/RR-96-449, November,1996 [3] Paredis J. Coevolutionary Computation. Artificial Life, 1995,2(4): 355-375 [4] Mitchbell A Potter. The Design and Analysis of a Computational Model of Cooperative Coevolution. PhD thesis, George Mason University,1997 [5] Daley, B, Schultz A, Grefenstette J. Co-evolution of Robot Behaviors. SPIE International Symposium on intelligent Systems and Advanced Manufacturing (ISAM '99), 19-22 Sept, 1999, Boston MA,1999 [6] David McFadzean. SimBioSys: A Class Framework for Biological Simulations. Master Thesis. The University of CalGary,1994 [7] Wooldridge M. Agent-based software engineering. IEE Proc. On Software Engineering, 144(1): 26-37 [8] 薛宏涛,冯庆堂,沈林成,常文森.多Agent系统的一个研究平台:基于SoccerServer的机器人足球赛仿真系统.计算机应用研究,1999.4 [9] 薛宏涛,沈林成,常文森.PTS领域中的Agent体系结构设计与实现.计算机工程与应用,2000.7 [10] 薛宏涛,叶媛媛,沈林成,常文森.多智能体系统体系结构及协调机制研究综述.机器人,已录用
