"CI" a particle swarm optimization algorithm with multi-scale cooperative mutation

"title" A Particle swarm optimization algorithm with multi-scale cooperative mutation (2010)

"thesis author" Tao Ning, Liufurong, Christina, Dong Zhijing

"Paper Link" Paper (14-pages)

Summary

This paper analyzes the influence of mutation operation on particle swarm optimization (PSO), proposes a new PSO algorithm with multi-scale cooperative mutation, aiming at the disadvantage of single mutation, slow convergence rate and easy to get into local minima, and proves that the algorithm takes probability 1 Convergence to the global optimal solution. This algorithm uses the multi-scale Gaussian mutation mechanism to realize the local solution escape. In the early stage of the algorithm, the large scale mutation and the uniform mutation operator can realize the fast localization of the global optimal solution space, with the increase of the adaptive value of the variation scale will be reduced, finally in the late stage of the algorithm, using PSO The evolutionary and small-scale mutation operators complete the search of local exact solution space, and effectively realize the organic coordination of exploration and mining ability. Applying the algorithm to 6 typical complex function optimization problems, and comparing with other PSO algorithms with mutation operation, the simulation results show that the algorithm not only has faster convergence speed, and the global solution search ability and stability are significantly improved.

"CI" a particle swarm optimization algorithm with multi-scale cooperative mutation