CAJ | 학술논문

对于函数优化问题，单种群遗传算法全局搜索能力较强，但局部搜索能力较弱，当函数为多峰时，易陷入局部解。本文引入多种群实数编码遗传算法，不同种群赋予不同的控制参数，以此兼顾全局和局部搜索能力，并采用移民算子联系各种群实现协同进化，以及人工选择算子保存各种群的最优个体作为终止判据，从而较好搜索到多峰函数的全局最优解，并提高迭代寻优效率。
대우함수우화문제，단충군유전산법전국수색능력교강，단국부수색능력교약，당함수위다봉시，역함입국부해。본문인입다충군실수편마유전산법，불동충군부여불동적공제삼수，이차겸고전국화국부수색능력，병채용이민산자련계각충군실현협동진화，이급인공선택산자보존각충군적최우개체작위종지판거，종이교호수색도다봉함수적전국최우해，병제고질대심우효솔。
The genetic algorithm has strong global searching ability, but its local searching ability is weak. This paper introduces multiple populations real coded genetic algorithm (MPRCGA) to optimize multimodal function. The different populations are endowed with different control parameters to guarantee both global and local search ability of MPRCGA. The immigration operator is used to contact multiple populations and realize co-evolution. The artificial selection operator is used to preserve the optimal individual as the criterion of termination in the evolution of multiple populations. The experimental results show that MPRCGA can not only solve the global minimum of the multimodal function, but also have high iterative optimization efficiency.