Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
496827 | Applied Soft Computing | 2009 | 10 Pages |
A particle swarm optimization (PSO) that uses an adaptive variable population size and periodic partial increasing or declining individuals in the form of ladder function is proposed in the paper. The aim is to enhance the overall performance of PSO. The proposed scheme adjusts the population size automatically according to the value of diversity of the population in ultimate time of current ladder. The processing of adding and declining the number of population is designed. The validity of the given algorithm is tested for a variety of benchmark problems and neural network training problems. The results of the proposed scheme are compared with the linearly decreasing inertia weight PSO (LDWPSO) and mutation PSO (MPSO), from which it is evident that the proposed scheme enhances the overall performance of PSO.