混沌差分进化算法在复杂优化问题中的应用研究  

Research on chaos differential evolution algorithm and its application to complex optimization problems

作　　者：肖文显[1] 许利军[2] 马孝琴[1] 

机构地区：[1]河南科技学院网络中心,河南新乡453003 [2]新乡学院网络中心,河南新乡453003

出　　处：《安徽大学学报（自然科学版）》2014年第3期32-36,共5页Journal of Anhui University(Natural Science Edition)

基　　金：国家自然科学基金资助项目(71171151);河南省教育厅自然科学基金资助项目(13B520011)

摘　　要：差分进化算法求解复杂优化问题时,由于进化后期种群多样性降低,算法极易陷入局部最优值无法跳出.论文针对该问题,将差分进化算法和混沌优化方法耦合,构建了混沌差分进化算法.算法利用混沌序列的遍历性和内部迭代的随机性,弥补差分进化算法容易陷入局部最优的缺陷,从而提高算法的搜索性能.对几种典型函数的测试结果表明:混沌差分进化算法的全局搜索性能有了显著提高,能有效避免算法陷入局部最优.因此,与标准差分进化算法和混沌优化算法相比,该算法在求解复杂优化问题时更加可行、有效.When differential evolution algorithm is used in solving the complex optimization problems, diversity of species is decreased in the later evolution period, therefore the algorithm can easily fall into local optimum. A novel chaos differential evolution algorithm based on the differential evolution and chaos optimization algorithm, which made use of the ergodicity and internal randomness of chaos iterations, was presented to overcome the defect of premature local optimum and enhance the global searching capacity of differential evolution with that of powerful local searching capacity of the chaos optimization algorithm. The experimental results indicated that the chaos differential evolution algorithm could improve the global searching capacity significantly and avoid falling into local optimum. Thus, the proposed approach was more feasible and effective in solving the complex optimization problem compared with differential evolution and chaos optimization algorithm.

关 键 词：复杂优化问题 遗传算法 混沌映射 混沌遗传算法 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]