Particle Swarm Optimization Algorithm in Transport Continuous Network Design Problems

We propose particle swarm optimization (PSO) algorithm for solving transport continuous network design problems (CNDP), and give sensitivity analysis for the parameters used in PSO. The CNDP is formulated as a bi-level programming model. The sensitivity analysis method, one-at-a-time designs (OATD), is used to analyze the effects of parameters. Numerical example demonstrates that PSO is an effective algorithm for solving CNDP with proper parameters setting. The choice of swarm size has a clear effect to the implementation time and with small swarm size may fail to the optimal solution. Furthermore, inertia weight and maximum velocity has clear effects to the solution searching.