Optimal Routing for Split Delivery Vehicle Routing Problems

Abstract

Split Delivery Vehicle Routing Problem (SDVRP) distributes the customers' demand to different vehicles for service. This study divides SDVRP into two stages for solutions. On stage one, Particle Swarm Optimization (PSO) was applied to separate customers into two bases, divisible and indivisible customer bases; and then another PSO was applied to resolve the Vehicle Routing Problem (VRP) of indivisible customers and a designed nearest insertion heuristic is utilized to insert the divisible customers into the proper location in the VRP. Meanwhile, a sweep algorithm with an initial heuristic rate are used to provide diverse PSO initial solutions so as to improve the efficiency on stage two. To verify the efficiency of the proposed method, SDVRP problem set given by J. M. Belenguer was applied for tests. The suggest methods came up with the experimental results shows that average deviations for the test instances are all within 10%. Restated, the proposed scheme of this work is able to resolve SDVRP optimization effectively.