| 多Agent协同设计系统学习机制 |
| |
| 引用本文： | 刘弘,郑向伟,王吉华.多Agent协同设计系统学习机制[J].兰州大学学报(自然科学版),2012,48(4):91-97. |
| |
| 作者姓名： | 刘弘 郑向伟 王吉华 |
| |
| 作者单位： | 山东师范大学信息科学与工程学院山东省分布式计算机软件新技术重点实验室,济南,250014 |
| |
| 基金项目： | 国家自然科学基金项目(60970004,60743010);教育部博士点基金项目(20093704110002);山东省自然科学基金项目(ZZ2008G02,ZR2010QL01);山东省重点实验室项目 |
| |
| 摘 要： | 从认知的和社会的角度分析了协同设计活动,提出了一种面向协同设计的多Agent系统结构和设计Agent的感知模型,以及多Agent协同强化学习的方法.该方法采用动态小生境技术对设计Agent进行分组,并选出每组中的最优设计Agent,使其通过与设计人员交互进行强化学习,然后和其他组选出的Agent协同学习,并把学到的知识在组内进行传播.以齿轮减速器设计为例,介绍了多Agent协同设计系统的协同设计及学习过程.
|
| 关 键 词： | 协同设计 多Agent系统 小生境技术 强化学习 |
Learning mechanism of a multi-agent cooperative design system |
| |
| LIU Hong
,
ZHENG Xiang-wei
,
WANG Ji-hua.Learning mechanism of a multi-agent cooperative design system[J].Journal of Lanzhou University(Natural Science),2012,48(4):91-97. |
| |
| Authors: | LIU Hong ZHENG Xiang-wei WANG Ji-hua |
| |
| Institution: | Key Laboratory for Distributed Computer Software Novel Technology of Shandong Province,School of Information Science and Engineering,Shandong Normal University,Jinan 250014,China |
| |
| Abstract: | Cooperative design activities were analyzed from cognitive and social viewpoints,and the architecture for a multi-agent system and a sensitive model of a design agent was put forward,thus presenting a multi-agent cooperative reinforcement learning approach for cooperative design.This approach adopts dynamic niche technology grouping design agents and selects the optimal design agent in every group.The selected agents can pursue reinforcement learning via an interaction with designers and carry on cooperative learning from each other,and then spread the learned knowledge in respective groups.A gear reducer design example was used to illustrate the cooperative design and learning process in a multi-agent cooperative design system. |
| |
| Keywords: | cooperative design multi-agent system niche technology reinforcement learning |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
|
