基于GE的多响应参数稳健设计优化方法  

Optimization method for robust parameter design in multi-response problems based on genetic evolution

作　　者：沈玲[1] 曾强[2] 杨龙飞[2] 张泽斌[2] 

机构地区：[1]河南理工大学安全科学与工程学院,河南焦作454000 [2]河南理工大学能源科学与工程学院,河南焦作454000

出　　处：《计算机工程与应用》2016年第5期250-254,共5页Computer Engineering and Applications

基　　金：河南省教育厅科学技术研究项目(No.12B120005);河南理工大学博士基金项目(No.B2011-088)

摘　　要：提出了一种基于遗传进化的多响应参数稳健设计优化方法。对多响应参数稳建设计问题进行了数学描述,建立了以试验样本多响应输出参数的加权平均质量损失最小化为目标的数学优化模型;提出了多响应参数稳健设计的遗传进化优化方法:以密集抽样取代离散化处理,以个体取代试验方案,以变化的种群取代固定的内表,通过遗传进化得到最优设计方案;提出并设计了多响应参数稳健设计的遗传优化算法。通过案例分析验证了该方法的有效性。An optimization method for robust parameter design in multi-response problems based on genetic evolution is proposed. The researched problem is described mathematically and an optimization model with the objective to minimize the weighted average quality loss of multi-response output parameters of experimental sample is bulit. An optimization method based on genetic evolution for the researched problem is presented. In the proposed method, the discretization is replaced by dense sampling, the experiment schemes are replaced by the individuals, the fixed internal table is replaced by transformable population, and the optimal design scheme is got through genetic evolution process. A genetic algorithm for robust parameter design in multi-response problems is presented and designed. The effectiveness of the proposed method is validated by case study.

关 键 词：参数稳健设计 多响应问题 遗传进化 遗传算法 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]