结合机械设计约束处理的差分进化算法  

Differential evolution algorithm with constraint handling in mechanical optimization design

作　　者：何兵[1,2] 车林仙[2] 刘初升[1] 

机构地区：[1]中国矿业大学机电工程学院,江苏徐州221008 [2]泸州职业技术学院机械工程系,四川泸州646005

出　　处：《机械设计》2012年第4期17-21,共5页Journal of Machine Design

基　　金：四川省应用基础研究计划资助项目(2008JY0163);泸州市重点科技计划基金资助项目(2010-S-21(2/7))

摘　　要：针对机械设计中的约束优化问题,提出了改进约束处理的自适应罚函数法。结合一般机械约束优化问题维数不高和差分进化算法简单、高效的特点,应用差分进化算法容易求得机械约束优化问题的全局最优解。给出了2个机械约束优化的数值实例,与已有的文献结果比较,表明新方法处理机械约束优化问题稳健且有效。An improved self-adaptive penalty function constraint handling method was proposed for the constrained optimization problem in the mechanical optimization design.In consideration of the features that the general mechanical optimization problem is not high in dimension and the differential evolution algorithm is simple and effective,the global optimal solution of mechanical constraint optimization problem can be easily obtained by applying the differential evolution algorithm.This paper gives two numerical examples for mechanical constrained optimization,compared with the results of the existing literature,indicating that the new approach is both effective and robust to solve mechanical constraint optimization problems.

关 键 词：机械设计 自适应罚函数法 约束处理 差分进化算法 

分 类 号：TH112.1[机械工程—机械设计及理论]