自适应爆炸半径的改进烟花算法

Improved fireworks algorithm with adaptive explosion radius

为充分利用被丢弃的爆炸火花个体的信息,对烟花算法进行优化,提出具有自适应爆炸半径特性的改进烟花算法。利用全局最优烟花个体gBest以及每个烟花所产生的最优爆炸火花个体的集合sparkpBest来构造新的爆炸半径,使其能够自适应地调整步长;在寻优过程中,对gBest进行高斯扰动来增加种群的多样性,避免烟花种群过快陷入局部最优。与其它群智能算法(粒子群算法PSO、带有高斯扰动的粒子群算法GPSO、蝙蝠算法BA、烟花算法FWA、自适应烟花算法AFWA以及增强烟花算法EFWA)对比,通过仿真可知,提出的改进烟花算法总体性能优于其它6种对比算法。

To make full use of the information of discarded explosive sparks,an improved firework algorithm called improved firework algorithm with adaptive explosive radius was proposed.An explosion radius was constructed to adjust the step size dyna-mically in the process of optimization by introducing the best global firework and the set of the best explosive spark produced by each firework.The Gaussian perturbation was carried out for best global firework to increase the diversity of population.Compared with other algorithms(particle swarm optimization,particle swarm optimization with Gaussian perturbation,bat algorithm,fireworks algorithm,adaptive fireworks algorithm and enhanced fireworks algorithm),the results show that the overall performance of the improved fireworks algorithm presented is better than the other six algorithms.