A Novel Self-Adaptation Hybrid Artificial Fish-Swarm Algorithm

Abstract

In this paper, the social behaviors of fish swarm were classified in four ways: foraging behavior, stray behavior, reproductive behavior, and escaping behavior. Inspired by this, a novel artificial fish swarm algorithm (NAFSA) was proposed, which integrated the merits of the self-adaptation strategy, mutation strategy and hybrid strategy into the social behaviors of fish swarm. In the case of mutation strategy, the cloud theory was introduced into the escaping behavior, and the basic cloud generator was used as the mutation operator because of the properties of randomness and stable tendency of a normal cloud model. For the hybrid strategy, the selection, crossover and mutation operator in evolutionary algorithm were applied to define the reproductive ability of an artificial fish. Furthermore, the parameters of Step and Visual were developed in forms of hyperbolic tangent function to adjust the optimize performance dynamically during iterations process. Finally, ten standard test functions are used as the benchmark to validate the effectiveness of the NAFSA. Experimental results confirmed the superiority of NAFSA in terms of both solution quality and convergence speed.