Particle Swarm Optimization with Adaptive Parameters

Particle swarm optimization is an effective evolution algorithm for global optimizing. Based on analysis of particle movements during evolution, parameter p is brought up to control the value of C₁ and C₂, which effects convergence rate of PSO. Aiming at solving different problems, corresponding p is adopted to improve performance. Particle confidence coefficient q is applied to weigh proper emphasize on itself best solution and global solution. Adaptive value of q is introduced to PSO to satisfy specific situation for each particle. Finally, performance of PSO with parameters p and q is testified by optimizing benchmark functions.