免疫逃避型粒子群优化算法

Particle Swarm Optimization based on Immune Evasion

针对基本粒子群优化算法容易陷入局部极值的缺陷,提出了一种免疫逃避型粒子群优化算法.其基本思想是将初始粒子群划分为寄生与宿主两个种群以模拟生物寄生行为,对寄生种群的粒子采用精英学习策略,对宿主群的粒子采用探索策略,再引入免疫系统的高频变异对寄生群采用相应的免疫逃避机制,以增强群体逃离局部极值、提高算法的全局寻优能力.采用标准测试函数的实验结果表明,该算法在收敛速度和求解精度方面均有显著改进.

A novel particle swarm optimization algorithm based on immune evasion （PSOIE） is proposed for conventional particle swarm optimization algorithms （PSO） often trapped in local optima. In order to improve the searching ability of the algorithm, we divided the particle swarm into two populations called parasitic and host. Elitist learning strategy is applied to the particle of parasitic population to avoid the elite particles into a local optimum. Exploration strategy and the clonal selection of immune system were introduced into host population to expand the search space of solutions and inhibit the premature stagnation. Immune evasion strategies were introduced into parasitic population to enhance the searching ability of parasitic population. The experimental results of six benchmark functions demonstrate the efficacy of the present algorithm.