A Grey wolf optimizer combined with Artificial fish swarm algorithm for engineering design problems

For the problems of Grey wolf optimizer (GWO) easy to fall into local optimum and lack of population diversity, this thesis raises a Grey wolf optimizer combined with an Artificial fish swarm algorithm (AFGWO). First, the search method of grey wolves is improved by combining the clustering behavior of Artificial fish swarm algorithm (AFSA) which avoids them falling into a local optimum. Second, to make the exploration and exploitation more balanced, the individual position of the worst grey wolf is updated by combining the foraging behavior of AFSA. Third, AFGWO adds quadratic interpolation and elite reverse learning to enrich population diversity. AFGWO is compared with other 10 algorithms on CEC2017 to evaluate its performance. The experimental results and four statistical analysis methods show that the proposed AFGWO based on above three improvement strategies has good solving ability and stability, and outperforms other comparative algorithms. In addition, AFGWO solves four classical engineering design problems well and demonstrates its good ability to solve realistic optimization problems. In general, AFGWO improved by the above strategy is better than GWO in performance, and is also very competitive with other intelligent algorithms.