解决一类特殊多目标优化问题的因子分析方法

根据国际最新提出的采用统计方法解决多目标优化的思想,提出了采用因子分析方法来解决多目标优化问题（FA-SMOEA）,在个体聚类过程中充分考虑了目标函数对pareto解的影响,用灵敏度向量和相似程度作为过渡量,将目标函数和个体聚类结合起来.在交叉和变异过程中再次考虑了公因子对个体进化的影响,利用公因子对个体进行了局部修正.最后用数值实验验证了其有效性.

Factor Analysis Method to Solve a Special Class of Multi-Objective Optimization Problem

According to the latest international statistical methods used to solve multi-objective optimization,this article proposes a factor analysis method to solve a special class of multi-objective optimization problem（FA-SMOEA）.This method,taking into account the objective function of the solution distribution in the process of clustering variables,cluster the objective function and the individual together in which sensitivity similarity are considered as transition amounts.Finally,it takes full account of the common factors on the evolution of the individual in the process of crossover and mutation.At the end of the article experimental results show its effectiveness.