Particle swarm optimization for the multi level capacitated minimum spanning tree

摘要

In the presented study Particle Swarm Optimization will be applied on an instance of the Multi Level Capacitated Minimum Spanning Tree Problem. Specifically a diversity preservation global variant of the PSO meta-heuristic will be presented. The particular PSO variant includes Gaussian mutation to avoid premature convergence and alternative selection of the flight guide per particle. Obtained results are compared with corresponding evolutionary approaches. Potential tree solutions are encoded/decoded using Network Random Keys. A real world network design case is introduced.