Guided Ejection Search for the Pickup and Delivery Problem with Time Windows

This paper presents an efficient route minimization heuristic for the pickup and delivery problem with time windows (PDPTW) based on guided ejection search (GES). GES is a recently proposed metaheuristc framework and was first applied to the vehicle routing problem with time windows. The existence of the pickup and delivery constraint makes the feasible solution space tightly constrained and then makes the design of effective metaheuristics more difficult. We demonstrate that GES can be successfully applied to such an complicated problem. Experimental results on the Li and Lim's benchmarks demonstrate that the proposed GES algorithm outperforms existing algorithms and is able to improve 105 best-known solutions out of 298 instances.