基于协方差分析的合作协同进化差分进化算法  

Cooperative coevolution algorithm with covariance analysis for differential evolution

作　　者：王彬[1,2] 任露[1] 王晓帆 曹雅娟[1] WANG Bin;REN Lu;WANG Xiaofan;CAO Yajuan(Faculty of Computer Science and Engineering,Xi’an University of Technology,Xi’an 710048,China;Shaanxi Key Laboratory for Network Computing and Security Technology,Xi’an University of Technology,Xi’an 710048,China)

机构地区：[1]西安理工大学计算机科学与工程学院,陕西西安710048 [2]西安理工大学陕西省网络计算与安全技术重点实验室,陕西西安710048

出　　处：《通信学报》2023年第1期189-199,共11页Journal on Communications

基　　金：国家自然科学基金资助项目(No.61976177,No.U21A20524)。

摘　　要：在大规模高维优化问题中,随着决策变量数目的增加,协同进化算法在搜索全局最优解过程中容易陷入局部最优。基于此,提出了一种基于协方差分析的合作协同进化差分进化算法,在根据决策变量之间的相关性对优化问题进行分组之后,针对子组件内部变量之间的相关性会影响种群进化过程的现象,在对子组件优化的过程中,利用协方差计算种群分布的特征向量,通过坐标旋转消除变量之间的相关性,有效避免在种群搜索过程中陷入局部最优,同时加快了算法的寻优速度。在CEC2014测试函数集上进行了对比实验,实验结果表明,所提算法具有可行性。With the increase of the number of decision variables, cooperative coevolution algorithm is easy to fall into local optimization in the process of searching the global optimal solution in large-scale high-dimensional optimization problems. Based on this, a cooperative coevolution algorithm with covariance analysis for differential evolution was proposed. After the optimization problems were grouped according to the correlation between the decision variables, the correlation between the internal variables of the subcomponents would affect the population evolution process. In the process of subcomponent optimization, covariance was used to calculate the characteristic vector of population distribution, and the correlation between variables was eliminated through coordinate rotation, which effectively avoided falling into local optimization in the process of population search and speeded up the optimization speed of the algorithm. Comparative experiments were carried out on the CEC 2014 test suite. The experimental results show that the proposed algorithm is feasible.

关 键 词：大规模优化问题 合作协同进化 相关性 协方差分析 差分进化 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]