论文部分内容阅读
针对合作协同进化算法(CCEA)动态适值空间的特点,研究信息补偿方法以消除由问题分解所导致的病态现象,并提出基于动态多种群进化策略的抗病态CCEA.每个协进化种群可动态分离出多个变化的子种群,利用它们同时获取多个全局或局部最优解作为交互信息,以实现信息补偿.针对引发病态行为的标准测试函数,与3种典型CCEA进行比较分析,实验结果表明所提出算法能有效克服病态现象,具有良好的全局优化能力.
Aiming at the characteristics of CCEA space, the information compensation method is studied to eliminate the pathological phenomena caused by the problem decomposition, and the CCEA based on dynamic multi-population evolution strategy is proposed. Each co-evolution population can be dynamically separated And use them to obtain multiple global or local optimal solutions simultaneously as the mutual information in order to compensate the information.According to the standard test function that caused the pathological behavior, compared with three typical CCEA, the experimental results show that The proposed algorithm can effectively overcome the pathological phenomena, has a good global optimization ability.