基于免疫原理的多目标进化算法群体多样性研究

RESEARCH ON POPULATION DIVERSITY OF MULTI-OBJECTIVE EVOLUTIONARY ALGORITHMS BASED ON IMMUNE PRINCIPLE

在多目标进化系统中一个关键问题是必须采取某一措施来保持群体多样性,本文将自然生命的免疫机制引入多目标进化过程,提出一种基于免疫原理的多目标进化算法群体多样性保持策略,并给出了具体设计方法,最后以两个目标的流水车间作业调度问题为例进行计算机模拟求解,将本文方法与"多目标遗传算法"的计算结果进行了对比,实验表明该方法能有效保持群体多样性,多目标优化结果的性能较好。

A key problem in a multi-objective evolutionary system is how to preserve diversity in the population. In the paper the mechanism of natural immune system is applied in multi-objective evolutionary process to solve this problem. A strategy of preserving population diversity of multi-objective evolutionary algorithms based on immune principle is introduced, and the detail design method is given. Finally, We describe the computer simulation of implementing the 2-objective flow shop scheduling problem, and compare the results of our method with that of Multi-Objective Genetic Algorithm. Experimental results show that this strategy can effectively preserve diversity in the population and it has better performances than traditional evolutionary approaches.