基于双极偏好占优的多目标进化算法及其应用

Multi-objective evolutionary algorithm based on bipolar preferences dominance and its application

为有效处理决策者能够提供双极偏好信息的多目标优化问题,加快原有算法的收敛速度,借鉴逼近理想解方法和搜索空间区域划分思想,定义了一种新型双极偏好占优关系,并引入到NSGA-Ⅱ算法中,设计了相应的非支配排序策略、种群多样性策略和约束处理策略,提出一种基于双极偏好占优的NSGA-Ⅱ算法(2p-NSGA-Ⅱ)。将该算法应用于求解两桁架结构设计的工程问题,对比仿真实验结果表明了2p-NSGA-Ⅱ算法的有效性。

To address the multi-objective optimization problems with bipolar preferences and to accelerate the conver gence speed of original algorithm, a new bipolar preference dominance relationship based on Technique for Order Preference by Similarity to an Ideal Solution （TOPSIS） and the idea of search space dividing was defined. By intro- ducing this relationship to Nondominated Sorting Genetic Algorithm Ⅱ （NSGA- Ⅱ ）, the corresponding strategies which included non-dominated-sorting procedure, population diversity and constraint-handling were designed. Thus a Bipolar Preference Dominance Based Nondominated Sorting Genetic Algorithm Ⅱ （2p-NSGA-Ⅱ ） was proposed. This algorithm was applied to the engineering problem of two bar truss design, the comparison result showed the ef- fectiveness of 2p-NSGA- Ⅱ.