基于免疫机理的多峰值函数并行优化算法

Multimodal Functions Parallel Optimization Algorithm Based on Immune Mechanism

多峰值函数的极值问题一直是优化领域中的一个难点和热点。传统搜索方法和遗传算法很难同时搜索出多个极值。基于生发中心中免疫细胞亲和度成熟的机理,本文提出了一种并行优化算法,目的是找出多峰值函数的多个最优解或最优解和尽可能多的局部优化解。算法的主要步骤有超变异、选择、记忆和相似性抑制。用不同的多峰值函数进行了仿真实验,并和相关算法进行了比较,结果表明所提出的算法具有良好的搜索性能。

Extremum problem of multimodal functions is a difficult issue in optimization fields. It is difficult for traditional search methods and simple genetic algorithm (SGA) to find out multi-local maximum simultaneously. A parallel optimization algorithm is proposed based on the idea of affinity maturation of immune cells in the germinal centers in order to achieve as many as possible local optimal solutions. The main steps of the algorithm include hyper-mutation, selection, memory and similarity suppression. The algorithm has been tested to optimize different multimodal functions, and the simulation results show that the algorithm is valid compared with other algorithms.