关键词: 人工鱼群算法, 随机移动算子, 粒子群算法, 自适应扰动

Abstract: By using basic artificial fish swarm algorithm to solve multi-peaks function optimization, there are many shortcomings, such as limited precision of calculation, partial optimal solutions only and easily, poor robustness, slower rate of convergence and the inefficiency of search, while the most important factor that causing those disadvantages is randomness of the random moved function. In order to ameliorate the moved function, this paper puts PSO algorithm and adaptive disturbance for random moved function into it, and then, Particle Swarm Optimization AFSA（PSO-AFSA） and Adaptive Disturbance Improved AFSA（ADI-AFSA） are proposed, what’s more, the convergence of proposed algorithm has been proved. Finally the results show that compared with AFSA and traditional AFSA, the two kinds of improved artificial fish swarm algorithms enhance the precision of calculation, convergence rates, the efficiency of search, and they have a better robustness by using a recognized set of functions which performs simulation tests.

Key words: Artificial Fish Swarm Algorithm（AFSA）, random moved function, Particle Swarm Optimization（PSO）, adaptive disturbance