Modified Dolphin Swarm Algorithm Based on Chaotic Maps for Solving High-Dimensional Function Optimization Problems

Abstract

In 2016, dolphin swarm algorithm (DSA) that has received sustained research interest due to its simplicity and effectiveness was proposed. However, when solving high-dimensional function optimization problems, DSA is prone to fall into local optimization problems, which leads to low optimization accuracy or even failure. In this paper, to solve this problem, chaotic mapping is introduced into DSA, and chaotic dolphin swarm algorithm (CDSA) is successfully proposed. Based on high-dimensional Rastrigin function, the optimal chaotic map is determined among eight chaotic maps (e.g., Logistic). Then, in view of high-dimensional Levy function, Rotated Hyper-Ellipsoid function and Sum Squares function respectively, the performance of CDSA and that of the state-of-the-art algorithms (e.g. (whale optimization algorithm) WOA) are compared. The results show that the performance of CDSA based on Kent map is best and the performance of CDSA outperform that of the state-of-the-art algorithms considered to be compared. Finally, it is concluded that such a new meta-heuristic algorithm could help to improve the shortcomings of DSA and increase the applied range of DSA.