A Swarm Intelligence Algorithm Based on Boundary Mutation

Abstract

In this paper, a new swarm intelligence algorithm called SIA for global optimization is proposed. Each individual in population is firstly projected onto the boundary of the search space in order to extend the region of the global search. Secondly, a probability of generating new individual is computed based on the function values of the individual, its corresponding boundary point and the best individual in the current population, then the new individual which is the variation of the individual is produced. These two steps are repeated until all individuals in the current population are processed. Then a new population is generated. The numerical experiments have indicated the efficiency of the proposed algorithm.