最大化个人偏好的多目标优化进化算法

针对多目标优化过程中如何将个人偏好信息融入寻优搜索过程的问题，本文提出一种最大化个人偏好以确定搜索方向的多目标优化进化算法．该算法首先采用权重和法将多目标问题转换为单目标问题，再利用遗传算法进行全局搜索，在满足个人偏好约束条件下，每一代进化结束后通过解约束优化问题获得能够使种群综合适应度具有最大方差的权重组合，从而最大化个人偏好以选择综合最优的个体进行遗传操作．按照不同个人偏好应用于传动系统进行控制器设计，仿真结果表明该算法能够获得满足个人偏好约束条件下的全局最优解．

The Evolutionary Multi-Objective Optimization Algorithm with Maximization of Personal Preference

For the problem that how to combine personal preference with the multi-objective optimization during the search procedure, an evolutionary multi-objective optimization algorithm which determines the search direction based on the maximization of personal preference is proposed. In the proposed algorithm, multi-objective problem is converted into a single objective problem by using the weighted-sum approach at first. After the conversion, genetic algorithm (GA) is used to search the feasible solution globally. Under the constraint of personal preference, the objective weights which make the synthetical finesses of the population own the biggest variance are calculated by solving a constraint optimization problem (COP). With maximization of personal preference, the optimal individual is selected exactly to execute the genetic operations. According to different preferences, the proposed algorithm is able to achieve the global optimal solution under the preference constraint.