ZHANG Junxi1, ZHANG Jiatong2， ZHANG Yumei3*
(1 Department of Vehicle Engineering, Xi′an Aeronautical University, Xi′an 710077, Shaanxi, China;2 College of Cultural Heritage, Northwest University, Xi′an 710069, Shaanxi, China;3 School of Computer Science, Shaanxi Normal University, Xi′an 710119, Shaanxi, China)
Abstract:
Based on the basic particle swarm optimization algorithm, the average of best particles is replaced by the weighted sum of the average of best particles and the average of the neighbor particles in the speed update formula. Using the ratio of average fitness of the whole particles and the average fitness of best particle as the fitness function, and the acceleration factor is introduced, a new adaptive PSO algorithm is obtained. The new algorithm used both the information of present particle and the information of the whole particles and the neighbors of present particle. In the process of evolution, it can adjust the global search and the local search component by the new model adaptively. So the convergence speed and precision can be improved. Experiments on 4 benchmark functions demonstrate that the new algorithm is more efficient.
KeyWords:
particle swarm optimization algorithm; fitness; update; convergence speed; convergence precision
