Cultured Artificial Fish Swarm Algorithm: An Experimental Evaluation

Abstract

In this work, a Weighted Cultural Artificial Fish Swarm Algorithm (wCAFSA) which is an amendment of standard artificial fish swarm algorithm (AFSA) is proposed. This algorithm can adaptively select its parameters at every generation in order to reduce the ease at which standard AFSA falls into local optimal. We first introduce inertial weight to adaptively determine visual distance and step size of AFSA thereafter, the Situational and Normative knowledge inherent in cultural algorithm are used to develop new variants of weighted cultural AFSA (wCAFSA Ns, wCAFSA sd, wCAFSA Ns+Sd and wCAFSA Ns+Nd). A collection of sixteen (16) optimization benchmark functions are used to test the performance of the algorithms. The simulation results disclosed that all the new variants of the wCAFSA outclassed the AFSA.