Skip to Main Content
The particle swarm optimization algorithm has been frequently employed to solve various optimization problems. Although the algorithm is performing satisfactorily while tackling unit-modal optimizations, enhancements in dealing with multi-modal functions are indeed desirable. Convergence of particles to the optimum solution is a primary and traditional requirement, however, this is achieved only after all the solutions space has been covered and evaluated. In this work, the focus is directed towards maintaining sufficient divergence of particles in multi-modal problems, by developing an alternative social interaction scheme among the swarm members. Particularly, a multiple-leaders strategy is employed in the new PSO algorithm to prevent pre-mature convergence. Results from benchmark problems are included to illustrate the effectiveness of the proposed method.