基于混合差分进化和alpha约束支配处理的多目标优化算法

Constrained multi-objective optimization with hybrid differential evolution and alpha constrained domination technique

针对约束多目标优化问题,提出了一种基于混合差分进化和alpha约束支配处理的优化算法.算法在用约束水平度对个体满足约束条件的程度进行定量化表达的同时融入支配关系.在初期放宽约束水平度,利用不可行解所携带的有用信息,增加种群多样性,在后期紧缩约束水平度,控制不可行解的比例,朝可行域方向进化.同时,将动态单纯形交叉算子和差分进化结合起来构成一种混合差分进化算法,提高算法的探索和开发能力.对6个典型测试函数求解的结果显示,本文算法无论是在收敛性方面还是解集分散性方面,与其它算法相比具有很大的优势.

To solve the constrained multi-objective optimization problems, we present a hybrid differential evolution algorithm with alpha constrained domination technique. In this approach, the constraint level, which measures how well an individual satisfies the constraints, is incorporated with the domination principle to solve multi-objective problems. At the early stage, the constraint level is relaxed in order to utilize the useful information carded by some infeasible individuals, so this relaxation increases the diversity of the population. At the later stage, the constraint level is tightened to make the evolution process searching for the feasible area. At the same time, a new dynamic simplex crossover operator is incorporated into differential evolution to improve the abilities of exploration and exploitation. The proposed algorithm is tested on 6 typical benchmarks and compared with other algorithms. Comparison results indicate that the proposed algorithm has advantages in converging to Pareto front and maintaining the evenly-distributed optima along the Pareto front.