摘要： 现有协作机制(CM)评价和选择模型存在通用性差、易陷入局部最优、CM库难以建立及评价函数设计困难等缺陷，不能满足现代复杂协同系统的要求。该文针对CM动态选择系统，构建CM评价的通用模型，设计对CM及其参数进行全局联合最优搜索的遗传算法，对模型和算法进行仿真。仿真结果表明，该模型能克服原有模型的缺陷，其算法能搜索到全局联合最优CM库。

关键词: 协作机制评价, 动态选择, 全局联合最优, 遗传算法

Abstract: The traditional Coordination Mechanism(CM) evaluation and selection models are not generic and prone to fall into local optimal decision. It is hard to establish a CM library and design evaluation function when using these traditional coordination mechanisms. The traditional coordination mechanisms cannot meet the needs of modern complicated cooperation systems. Aiming at the CM dynamic selection system, a generic model of CM evaluation is proposed, and a GA of search for global joint optimal CM and its parameter is designed. The generic model and genetic method are simulated. Simulation results show that this model overcomes the traditional models’ drawbacks, and its algorithm can search out the global joint optimal CM.

Key words: coordination mechanism evaluation, dynamic selection, global joint optimal, Genetic Algorithm(GA)

中图分类号: 

• TP302