求解约束优化问题的动态目标迁移差分进化算法  

An Dynamic Objective Migration Differential Evolution Algorithm of Constrained Optimization Problems

作　　者：刘俊梅[1] 马永刚[1] 高岳林[2] 

机构地区：[1]中国矿业大学银川学院基础部数学教研室,银川750011 [2]北方民族大学信息与系统科学研究所,银川750021

出　　处：《武汉理工大学学报（交通科学与工程版）》2012年第5期1083-1087,共5页Journal of Wuhan University of Technology（Transportation Science & Engineering）

基　　金：国家自然科学基金项目资助(批准号:60962006)

摘　　要：针对约束优化问题的特点,给出一种改进差分进化算法.将差分进化算法做了适当修正,在初始化中加入迁移操作,依据违反约束度函数和原目标函数进行选择操作,当个体的违反约束度在容忍度以外时,通过违反约束度函数更新个体,当个体的违反约束度在容忍度以内时,通过原目标函数更新个体,实验研究结果表明该算法能有效求解约束优化问题.For constrained optimization problem, an improved differential evolution algorithm is given. In the algorithm, differential evolution algorithm is appropriately modified. Migration operation is used in initialize. According to constraint violations function and original objective function selected operate. When the individual degree of violation constraints beyond the tolerance, the individual upda- ted by constraint violations function. When the individual degree of violation constraints within the tolerance, the individual updated by objective function. Numerical simulations show that this algo- rithm is effective for solving constrained optimization problems.

关 键 词：约束优化 差分进化算法 动态目标 迁移操作 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]