An Improved Particle Swarm Optimization Algorithm for High Dimensional Multimodal Optimization Problems

摘要

An improved particle swarm optimization is proposed in this paper in order to solve the optimization of multimodal functions with high dimensions and overcome the shortcoming of only one extremum can be found in standard particle swarm optimization. The evolvement of particles is divided into two stages: the phase of self-study and of global-study. In the first phase, particles explore the potential extremums alone through mutation and at the same time the strategy of local updating is adopted to receive the good mutations. In order to find the global best position and speed up convergence, the produce of the global best particle is modified in the second phase so that the global best particle can fly following the best position of each dimension. The typical numerical simulation results show that the improved algorithm is fairly effective.