一种基于免疫原理的多种群DNA遗传算法

A muti-population DNA genetic algorithm based on immune principle

针对单种群遗传算法易陷入局部最优、多样性丧失快等问题,提出一种基于免疫原理的多种群DNA遗传算法。在多种群协同进化的基础上,将DNA计算思想引入到编码和遗传操作算子的设计中,通过模拟生物机体的免疫机制对遗传进化过程中个体的产生和选择过程进行自适应调控,并利用优良个体的迁移实现种群间信息交流。最后,通过函数优化实验测试算法的性能。仿真结果表明,算法在发掘全局最优个体、局部搜索能力方面表现优越。

In view of the disadvantages of easily trapping in local optimum, rapid diversity loss of single population genetic algorithm, an immune principle-based muff-population DNA genetic algo- rithm was proposed. Based on multi-population co-evolution, DNA computing was introduced for parameter coding as well as the design of genetic operators. Generation and selection of individual in the process of genetic evolution was adaptively regulated by imitating immune mechanism of living organisms, communication of populations was realized by the migration of excellent individu- als. Performance of the algorithm was tested by function optimization. The simulation results show that the algorithm is superior in the ability of finding global optimum as well as local searching capability.