Improved Particle Swarm Optimization for Minimum Spanning Tree of Length Constraint Problem

Minimum spanning tree of length constraint problem (MSTLCP) can not be solved by traditional algorithms, so improved Particle Swarm Optimization (PSO) is put forward to solve MSTLCP, based on the idea of global and feasible searching. In the improved PSO, introducing check function ensures the feasibility of particle swarm; more reasonable fitness function is designed based on the relationship between spanning tree and its cotree; improvements for updating the position of particle make the current position of particle be beneficial to being close to the best position of particle in its neighborhood. Therefore improved PSO on MSTLCP is more reasonable than traditional PSO on MSTLCP. And then by doing simulation experiment, the analysis of the change of the parameters and the analysis of the change of particle swarm scale and iteration number, we draw a conclusion that improved PSO is a practical algorithm.