首页 | 本学科首页   官方微博 | 高级检索  
     

基于历史行为的agent联盟策略
引用本文:蒋建国,尹翔,夏娜,苏兆品.基于历史行为的agent联盟策略[J].电子学报,2007,35(8):1485-1489.
作者姓名:蒋建国  尹翔  夏娜  苏兆品
作者单位:合肥工业大学计算机与信息学院,安徽合肥 230009
基金项目:国家自然科学基金,高等学校博士学科点专项科研项目,安徽省自然科学基金
摘    要:协商是多agent系统中联盟形成的重要手段,很少有对协商初始提议的讨论.提出了一种基于个体历史行为确定协商起点的方法,根据agent在历史任务中的各维能力贡献率,计算其在当前任务中的期望获利因子,并据此给出协商时的初始提议,在此基础上设计了一种联盟形成策略,试验表明了该方法的有效性.

关 键 词:初始提议  能力贡献率  期望获利因子  历史行为  
文章编号:0372-2112(2007)08-1485-05
收稿时间:2006-08-04
修稿时间:2006-08-04

Strategy of Agent Coalition Based on Historic Behavior
JIANG Jian-guo,YIN Xiang,XIA Na,SU Zhao-pin.Strategy of Agent Coalition Based on Historic Behavior[J].Acta Electronica Sinica,2007,35(8):1485-1489.
Authors:JIANG Jian-guo  YIN Xiang  XIA Na  SU Zhao-pin
Affiliation:Department of Computer and Information Science,Hefei University of Technology,Hefei,Anhui 230009
Abstract:Negotiation is an important method to coalition formation in MAS.Most of existing work seldom studies how to generate initial proposals.A strategy based on historic behavior to decide initial proposals is proposed.By evaluating each agent's contribution ratio in historic tasks,the expected profit factor to the current job can be attained and based on it the initial proposal is obtained.Moreover,a coalition formation strategy is presented.Experimental results demonstrate that this method is efficient.
Keywords:initial proposals  contribution ratio in capabilities  expected profit factor  historic behavior
本文献已被 CNKI 维普 万方数据 等数据库收录!
点击此处可从《电子学报》浏览原始摘要信息
点击此处可从《电子学报》下载全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号