CAJ | 학술논문

提出一种择优学习的多个体差分算法用于求解约束优化问题，目的是用来提高差分算法的搜索能力。首先，将择优学习策略应用到混合变异算子中，使其快速搜索到可行解区域，然后使用克隆策略加大对最优解区域的搜索力度，增强局部搜索能力。通过对CEC2006经典Benchmark函数测试，实验结果表明本算法在求解效率和精度上都取得了较好的结果。
제출일충택우학습적다개체차분산법용우구해약속우화문제，목적시용래제고차분산법적수색능력。수선，장택우학습책략응용도혼합변이산자중，사기쾌속수색도가행해구역，연후사용극륭책략가대대최우해구역적수색력도，증강국부수색능력。통과대CEC2006경전Benchmark함수측시，실험결과표명본산법재구해효솔화정도상도취득료교호적결과。
This paper presents a preferred learning-based multiple individuals differential evolution algorithm for solving constrain-ed optimization problems, in order to improve the search ability of difference evolution algorithm.Firstly, a kind of preferred learning strategy is applied to the hybrid mutation operator, which makes it quickly search to the area of feasible solution, then the clone strategies are used to enhance the search intensity in optimal solution area and the ability of local search is greatly im-proved.Finally the algorithm is tested on the CEC2006 classic Benchmark function, the experimental results show that the algo-rithm is of better results in solving efficiency and accuracy.