Development and analysis of a modified Artificial Fish Swarm Algorithm

The analytical development and simulation of a modified Artificial Fish Swarm Algorithm is presented in this paper. It addresses the problem of the conventional AFSA migration into the local minima due to the constant effects of its control parameters such as the visual distance and step size. An iterative behavior was introduced such that each generation can select its guiding parameters adaptively. The performance of the modified algorithm was evaluated using a subset of applied mathematical optimization test functions namely Ackley, Cosine Mixture, Neumaier, Rosenbrock and Rastragin. Results show that the modified algorithm performed better in four (4) out of the 5 test functions.