Abstract: By utilizing the immunity theory, the immunity operator can restrain the degenerate phenomenon during the evolutionary process. This paper designs a novel immunity genetic algorithm: the “blind mutation operator” is substituted by the “guided mutation operator”, which is immunity operator. By changing the probabilities of crossover and immunity, it examines the role of crossover and immunity operator in immunity- evolutionary algorithm with the examples of traveling salesman problem. The experimental results validate the constructive aspect of crossover and local search ability of immunity operator to restrain the degenerate phenomenon of crossover.

Key words: algorithm design, crossover operator, immunity operator, traveling salesman problem

摘要： 通过应用免疫原理设计免疫算子对交叉结果进行修复，将免疫算子作为“有导向的变异算子”取代经典演化算法中的“盲目的变异算子”，可以有效抑制优化过程中的退化现象。该文以旅行商问题为例，通过设置不同的交叉概率和免疫概率，对免疫遗传算法中交叉算子和免疫算子的作用进行研究。实验结果表明，交叉算子的作用在于保证候选解的多样性和建设性（将局部近似最优解组合成全局近似最优解）；而免疫算子的作用则是对候选解进行局部优化，二者的结合体现了搜索过程中勘探与开采的统一。

关键词: 算法设计, 交叉算子, 免疫算子, 旅行商问题

CLC Number: 

• TP301.6