基于差异演化算法和多属性决策的机电系统可靠性多目标优化设计

针对机电系统可靠性设计问题，以可靠性和费用(或体积等)最优为目标建立可靠性设计的多目标优化模型．提出了自适应多目标差异演化算法，该算法提出了自适应缩放因子和混沌交叉率，采用改进的快速排序方法构造Pareto最优解，采用NSGA-II的拥挤操作对档案文件进行消减．采用自适应多目标差异演化算法获得多目标问题的Pareto最优解，利用TOPSIS方法对Pareto最优解进行多属性决策．实际工程结果表明：自适应多目标差异演化算法调节参数更少，且求得的Pareto最优解分布均匀；采用基于TOPSIS的多属性决策方法得到的结果合理可行．

Mechanical electronics system reliability multi-objective optimization design based on differential evolution algorithm and multiple attribute decision making

In order to improve mechanical electronics system reliability， a multi-objective optimization model based on reliability indices and cost (volume et al.) was constructed. Adaptive Multi-Objective Differential Evolution Algorithm (AMODE) was proposed. The algorithm proposed chaotic crossover ratio and adaptive scaling factor， improved quick sort algorithm was employed to build Pareto optimal solutions， and the shrink of an external archive was achieved based on crowding distance， which was used in NSGA-II. The Pareto optimal solutions were obtained by AMODE. Multiple attribute decision making method TOPSIS was adopted to rank these solutions. An actual project solution shows that less parameters of AMODE need adjust and the algorithm can obtain uniformly distributed Pareto optimal solutions. The result based on multiple attribute decision making method-TOPSIS is reasonable and feasible．