Discrete Particle Swarm Optimization for Solving a Single to Multiple Destinations in Evacuation Planning

摘要

During the evacuation process, the most challenging task is to move people to safer locations. As time is the decision factor in the evacuation process, urgent and firmly decisions are required. An evacuation plan should be efficiently constructed by taking into account the routes of vehicle. This paper explored the feasibility of using a discrete particle swarm optimization (DPSO) algorithm to solve evacuation vehicle routing problem (EVRP) problem focusing on a static vehicle routing. A linear mathematical formulation is constructed with the objective to find a minimum total travelling of the capacitated vehicles from vehicle location to various number flooded areas. A solution representation based on a search decomposition procedure is proposed to accommodate the routing process and mapped to the discrete multi-valued particle in DPSO. Computational experiment involves datasets from the event of flash flood. Comparative analyses were carried on both DPSO and GA. The results indicate that the DPSO are highly competitive and showed good performance in both fitness value (total travelling time) and processing time.