基于多种群差分进化的多目标优化算法

摘要

In order to avoid the situation of falling into local optimum in solving the multi-objective optimization prob-lem(MOP) with differential evolution algorithm(DE), we designed a bidirectional search mechanism which can improve the ability of local search of the DE. We also designed a multi-population mechanism for DE, which can reduce the risk of local optimum, and make the Pareto fronts more evenly distributed. Experimental results shows that, compared with similar algorithms such as NSGA-II,the proposed method is more efficient, while the precision and distribution of Pare-to optimal solution set is better than the former.%针对差分进化算法(Differential Evolution Algorithm,DE)求解多目标优化问题时易陷入局部最优的问题,设计了一种双向搜索机制,它通过对相反进化方向产生的两个子代个体进行评价,来增强DE算法的局部搜索能力；设计了多种群机制,它可令各子群独立进化一定次数再执行全局进化,以完成子群间进化信息的交流,这一方面降低了算法陷入局部最优的风险,另一方面增强了Pareto解集的多样性,使Pareto前沿面的解集分布更为均匀.实验结果表明,相比于NSGA-Ⅱ等同类算法,所提方法在搜索Pareto最优解时效率更高,并且Pareto最优解集的精度及分布程度比前者更好.