基于空间自适应收缩策略的混合遗传算法及其应用  

Hybrid Genetic Algorithm Based on the Strategy of Searching-Space Adaptive Shortening and Its Application

作　　者：李玲玲[1] 郭素娜[1] 袁绍民[1] 崔志华[2] 

机构地区：[1]河北工业大学电气与自动化学院,天津300130 [2]太原重型机械学院系统仿真与计算机应用研究所,山西太原030024

出　　处：《低压电器》2006年第9期3-7,39,共6页Low Voltage Apparatus

基　　金：河北省博士基金资助项目(05547003D-3)

摘　　要：将简单遗传算法与单纯形法相杂合,设计了一种实数编码的混合遗传算法(HGA),用于求解无约束优化问题。算法采用了最优保留策略,同时在变异操作中采用了搜索空间的自适应收缩策略,以提高全局和局部搜索能力,加快收敛速度,避免退化。在求解约束优化问题时,利用罚函数处理约束条件,由HGA对增广目标函数寻优。HGA的有效性通过3个典型测试函数得到验证,并应用于拍合式继电器电磁系统的体积优化。A real-coded hybrid genetic algorithm （HGA） was designed by hybridizing the simple genetic algorithm and simplex algorithm to solve unconstrained optimization problems. In order to avoid degeneration and to improve the global and local searching property and the convergence speed, the optimum maintaining strategy was adopted and the searching-space was shortened adaptively via mutation operator. When solving a constrained optimization problem, a set of penalty functions was used to process the constraints and HGA was used to optimize the generalized objective function. The validity of HGA in this work was verified by three typical functions. Further more, an application was introduced for the volume optimization of electromagnet system of clapper-type relay.

关 键 词：混合遗传算法 单纯形法 搜索空间 自适应收缩 罚函数 约束优化问题 

分 类 号：TB114.3[理学—概率论与数理统计]