Comparative Study of Horse Swarm Algorithm and Classical Algorithm

Abstract

At present, optimization algorithms have been widely used in various scientific fields. These optimization algorithms are usually stimulated by the natural behavior of human, animal, plant, physical or chemical reagents. Most of the algorithms proposed in the past decade are inspired by animal behavior. Based on the horse swarm optimization algorithm, a horse swarm algorithm (WHO), which simulates the social life behavior of horses, is proposed in this paper. The new algorithm integrates the golden sinusoidal guiding mechanism as local operators into WHO algorithm, which improves the accuracy and convergence speed of the algorithm; It avoids the early over convergence of the algorithm. On the challenging CEC2019 test set, the WHO algorithm is comprehensively compared with other improved algorithms. Simulation results show that WHO algorithm has better performance in search efficiency, convergence accuracy and avoiding local optimum for both high-dimensional and fixed-dimensional problems. The results show that compared with other algorithms, this algorithm has strong competitiveness.