双目标动态迁移操作DE-PSO混合算法  

Dual target dynamic migration DE-PSO hybrid algorithm

作　　者：赵华[1] 刘俊梅[2] 

机构地区：[1]北方民族大学文史学院,宁夏银川750021 [2]中国矿业大学银川学院基础部数学教研室,宁夏银川750011

出　　处：《宁夏工程技术》2013年第4期315-318,共4页Ningxia Engineering Technology

摘　　要：针对非线性约束优化问题的特殊性,给出一种求解非线性约束优化问题的动态目标迁移DEPSO混合算法.在初始化中加入迁移操作,采取动态目标的处理方法,将约束优化问题转化为无约束双目标优化问题.依据原目标函数、违反约束度函数进行选择操作,先通过改进差分进化算法对种群进化,对违反约束度在容忍度以外的个体再采用改进的粒子群优化算法进化,并用采用一组经典的测试函数进行测试.DE-PSO混合算法具有精度高、稳定性好的特点.For the particularity of nonlinear constrained optimization problems, a dynamic target migration DE-PSO hybrid algorithm is given to solve the nonlinear constrained optimization problem. Adding migration to the initialization operation, taking dynamic target processing method, the constrained optimization problem is changed into an unconstrained bi-objective optimization problem. According the original objective function and constraint violations degree function, selection operation is executed. Using improved differential evolution algorithm to solve the initial population, and using improved particle swarm optimization algorithm, the individual which is violation to constraint value and out of tolerance uses improved PSO evolution. All the experiments are tested by the classical trial functions. Results show that the algorithm has the higher precision and better stability than the other two.

关 键 词：差分进化(DE) 动态目标迁移 粒子群优化(PSO) 非线性约束优化问题 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]