AM模式下车间多智能体协调研究  

Research on multi-agent coordination of workshop in AM pattern

作　　者：张铭华[1] 苏锦旗[1] 

机构地区：[1]西安邮电大学管理工程学院,陕西西安710061

出　　处：《武汉大学学报（工学版）》2015年第1期130-137,共8页Engineering Journal of Wuhan University

基　　金：国家自然科学基金项目(编号:71102149);国家社会科学基金项目(编号:11CJY064);工信部通信软科学研究项目(编号:2013R01-2);教育部人文社会科学研究项目(编号:12YJC790084);陕西省教育厅专项科研计划资助项目(编号:12JK0056);西安邮电学院青年教师科研基金项目(编号:ZL2011-22)

摘　　要：为满足现代制造业的敏捷性要求,面向敏捷车间的单元动态协调方案设计开发,构建敏捷车间多智能体动态协调框架.基于文献述评,提出自上至下以及自下至上相结合的智能体动态协调思想,分析智能体动态协调的内涵及特征,建立融入模糊理论的智能体动态协调综合满意度函数.考虑敏捷车间中Agent的有限理性以及决策支持信息的不完全不完美性,运用演化博弈论,给出Agent的资源优选演化稳定策略解释.接着,利用协同演化遗传算法求得面向多项工作任务的车间资源重组全局优化解,结合聚类分析法作出敏捷车间智能体动态协调的最满意决策.最后提供陕西装备制造业实例,验证了上述协调算法的有效性及可行性.In order to meet agility of modern manufacturing industry, unit dynamic coordination framework of agile workshop was built based on multi-agent system, facing unit dynamic coordination solution design and development of agile workshop. Based on literature review, dynamic coordination thought was pro- posed with combination of both up-bottom and bottom-up approach. Connotation and characteristics of unit dynamic coordination were also analyzed. Then, with integration of fuzzy theory, comprehensive satisfac- tion degree function of unit dynamic coordination was established. In consideration of limited rationality, incompletion and imperfection of Job Agent decision support information in agile workshop, evolutionary stable strategy of resources optimization was explained based on evolutionary game theory. By co-evolution genetic algorithm, overall optimization solution of workshop resources reorganization was acquired for many jobs; and the most satisfied decision of unit dynamic coordination in agile workshop was made with combination of clustering analysis method. Finally, coordination example in Shaanxi equipment manufac- turing industry was provided; and its validity and feasibility were proved,

关 键 词：敏捷车间 动态协调 多智能体系统 演化博弈 协同演化遗传算法 聚类分析 

分 类 号：TH166[机械工程—机械制造及自动化] TP18[自动化与计算机技术—控制理论与控制工程]