An improved chaotic artificial fish swarm algorithm and its application in optimizing cascade hydropower stations

As a newly-proposed stochastic global optimization algorithm, artificial fish swarm algorithm (AFSA) is featured by its good global convergence and high convergence speed. However, it may suffer from the problem of being trapped in local optimum and it has relatively low search accuracy. Having analyzed the deficiencies of AFSA and making use of the ergodicity and internal randomness of chaos optimization algorithm (COA), this paper further puts forward an improved chaotic artificial fish swarm algorithm (ICAFSA). In this improved algorithm, chaos optimization is first employed to initialize the position of individual artificial fish and then AFSA is applied to obtain the neighborhood of the global optimum solution. When there is no change or little change of the function values on bulletin board in successive iterations, chaotic mutation is then executed to help the artificial fish swarm get rid of the local optimum. The findings of case study show the feasibility and effectiveness of the ICAFSA in the optimization operations of cascade hydropower stations.