摘要
本文首先分析了多Agent协同技术的发展现状和趋势，对多Agent协同中的启发式任务分配算法、基于排队论和基于Tribase模型的任务分配算法进行了介绍，并对在Agent通信领域占主导地位的两种通信语言——KQML和FIPA-ACL以及黑板方式和点对点方式等常见的Agent通信连接方式进行了探讨。在对任务分配和多Agent通信管理等问题进行较为深入的讨论的基础上，本文分析了几个典型的多Agent系统结构，并将多Agent协同技术引入企业信息管理领域，提出了基于冗余服务管理的多Agent协同模型——RSMACM。RSMACM模型结合服务队列理论以及冗余服务管理技术，采用XML语言进行通信消息的表达，对负载均衡、通信效率和可靠性、复杂语义通信及其灵活性等方面进行了研究和实践。文章最后结合南京禄口国际机场生产运营管理信息系统项目，阐述了RSMACM模型的应用与实现，并对实际的运行效果进行了分析评价。
This paper first analyses the development status and tendency of multi-agent cooperation technology, and introduces a heuristic task distribution algorithm, the algorithms based on queuing theory and the algorithm based on tribase model in multi-agent cooperation. Then this paper discusses KQML and FIPA-ACL as two leading agent communication languages and analyses usual connecting methods in communication among the agents such as blackboard and point-to-point method. Then based on the discussion of task distribution and communication among the agents, this paper analyses architecture of some typical multi-agent systems, then introduces multi-agent cooperation technology into the field of enterprise information management, and brings forward the Redundant Service Management Based Multi-Agent Cooperation
Model (RSMACM). The model RSMACM integrates queuing theory and redundant service management technology, and utilizes XML language to express messages, and according to which, this paper implements and researches on load balance, efficiency and reliability in agent communication, complex semantic communication and its flexibility. Finally, based on Nanjing Lukou International Airport's Management Information System (LKMIS), the paper discusses the application and implementation of RSMACM, and gives the evaluation of the system's performance.
引文
[1]袁贞明，庄越挺．基于多Agent系统的仿真系统研究．模式识别与人工智能．2001 Vol．14 No．3：266—271
[2]赵文龙，侯义斌．多Agent系统的组织结构与协同．计算机工程与应用．2000 No．10：59—61
[3]赵文龙，侯义斌，多Agent系统及其组织结构，计算机应用研究．2000 No．7：12—14
[4]赵文龙，侯义斌．合作Agent的构造问题．计算机应用研究．2002 No．1：4—7
[5]何炎祥，宋文欣，彭锋，陈莘萌．MDOCEM中的一种启发式任务分解算法．小型微型计算机系统．1999 Vol．20 No．12：893—895
[6]陈进才，郑守淇，冀德学．Agent协作中的分布渐进协商冲突消除策略．微电子学与计算机．2000 No．1：5—9
[7]魏晓斌．一个基于XML的Agent通信框架．计算机应用研究．2001 No．10：108—112
[8]陈宁江，苏德富．基于Agent的CSCW系统及其通讯实现，广西大学学报(自然科学版)．2000 Vol．25 No．2：94—97
[9]刘向军，刘世平，张洁，李培根．多Agent系统通信与协作机制构造．机械设计与制造工程，2002 Vol．31 No．2：40—42
[10]鲍钟峻，张琼，石教英．采用多Agent模型的协同编著系统的设计与实现．计算机研究与发展，1998 Vol．35 No．11：966—970
[11]王寻羽，朱淼良，邱瑜，唐文斌．多Agent系统中实时消息传递机制．计算机科学．2000 Vol．27 No．6：50—54
[12]王继宏，胡建平．有限资源环境下的分层、分布式体系结构研究[J]．计算机科学．2000 Vol．27 No．3：26-28
[13]钱方等．分布式计算环境中面向冗余服务的对象模型[C]．全国开放式分布与并行处理学术会议论文集．1999：128-133
[14]何炎祥，吴思，宋文欣，彭锋，陈燕涛．MDOCEM中的可重用资源管理策略．小型微型计算机系统，1998 Vol．19 No．9：12—17
[15]苑迎春，王凤先，田俊峰．基于多Agent的冗余服务管理研究．计算机应用研究．2001 No．11：44—46
[16]胡舜耕，刘晓宇，钟义信，王克宏．面向自动文摘的多Agent系统中的协调算法研究．计算机研究与发展．2001 Vol．38 No．11：1302—1309
[17]胡代平，王浣尘．预测模型Agent的实现方法．系统工程学报．2001 Vol．16 No．5：330—334
[18] 《运筹学》教材编写组．运筹学．北京．清华大学出版社．1986：312—314
[19] 叶华，张森．一种异构数据库间的XML接口方案．计算机应用．2001 Vol．21 No．6：15—17
[20] 黄理，曹林有等．全面引爆XSP网站开发——XSP／Cocoon／XML核心技术内幕[M]．北京希望电子出版社．2002：2—10
[21] Agent Name Server. http://ailab.im.fju.edu.tw/camase/ans/default.asp
[22] Katia P Sycara. Multiagent Systems. AI Magazine. 1998 Vol.19 No.2:79-92
[23] Katia P Sycara. Intelligent Agents and the Information Revolution. UNICOM Seminar on Intelligent Agents and their Business Applications. London. 1995: 143-159
[24] The Pleiades:http://www. cs.cmu.edu/afs/cs.cmu.edu/project/theo-5/www/pleiades.htm
[25] Terry R. Payne, Rahul Singh, Katia P Sycara. Calendar Agents on the Semantic Web.IEEE Intelligent Systems. June 2002:84-86
[26] Katia P Sycara, K. Decker, A. Pannu, M. Williamson, D. Zeng. Distributed Intelligent Agents. IEEE Expert-Intelligent Systems and Their Applications. 1996 Vol. 11 No.6: 36-45
[27] Katia P Sycara, K. Decker, A. Pannu, M. Williamson. Designing behaviors for information agents. Proceeding of Agents-97. Los Angeles. 1997:404-412
[28] Dwight Deugo, Michael Weiss, Elizabeth Kendall. Reusable Patterns for Agent Coordination. http://bim.im.fju.edu.tw/home/
[29] G. Cornuejols, G. Nembouser, L. Wolsey. The Uncapacited Facility Location Problem.Carnegie-Mellon University, MSRR. 1983: 493-510
[30] Viadimir Marik, Michal Pechoucek, Olga Stepankova, Jiri Lazansky. Proplant: Multiagent System for Production Planning. Applied Artificial Intelligence. 2000 Vol. 14:727-762
[31] Jennings N. R., T. Wittig. ARCHON: Theory and practice. In distributed AI: Theory and Praxis, eds. M. Avouris and L. Gasser. Dordrecht: Kluwer Academic Publishers. 1993:179-195
[32] Nwana H. S., D. T. Ndumu. An Introduction to Agent Technology. Software Agents and Softcomputing, eds. H. S. Nwana and N. Azarmi, LNAI 1198. Heidelberg: Springer-Verlag. 1997:3-26
[33] Cao W., C.-G. Bian, G. Hartvigsen. Cooperator-base and Task-base for Agent Modeling: The Virtual Secretary Approach. Pro. of AAAI-96 Workshop on Agent Modeling. Menlo Park, CA: AAAI Press. 1996:105-111
[34] Fundation for Intelligent Physical Agents. FIPA 97 Specification Part 2: Agent Communication Language[M]. http://www.fipa.org/
[35] F. Dignum. Information Management at a Bank Using Agents: Theory and Practice.
Applied Artificial Intelligence. 2000 Vol. 14:677-696
[36] M Wooldridge. Verifiable Semantics for Agent Communication Languages[C]. 3rd Int. Conf. on Multi-Agent Systems (ICMAS'98). IEEE Computer Society Press. 1998
[37] Sycara K., D. Zeng. Coordination of Multiple Intelligent Software Agents.International Journal of Cooperative Information Systems. 1996 Vol.5 No.2-3: 1-31
[38] Katia Sycara, Massimo Paolucci, Martin van Velsen, Joseph Giampapa. The RETSINA MAS Infrastructrue. http://www. cs.cmu.edu/～softwareagents
[39] Onn M. Shehory, Katia Sycara, Somesh Jha. Multi-agent Coordination through Coalition Formation. http://www. cs.cmu.edu/～softwareagents
[40] Wu Wen, Fumio Mizoguchi. An Authorization-based Trust Model for Multiagent Systems. Applied Artificial Intelligence. 2000 Vol. 14:909-925
[41] Extensible Markup Language (XML)[EB/OL]. http://www. w3.org/xml/
[42] E. R. Harold. XML Bible[M]. IDG Books Worldwide, Inc. 1999:1-5
[43] The DARPA Knowledge Sharing Initiative External Interfaces Working Group.DRAFT Specification of the KQML Agent-Communication Language. http://ailab.im.fju.edu.tw/kqml/