A hybrid large neighborhood search for dynamic vehicle routing problem with time deadline

Abstract

In order to consider the customer’s real-time requests and service time deadline, this paper studies the dynamic vehicle routing problem with time deadline (DVRPTD), and proposes a hybrid large neighborhood search (HLNS) algorithm. This strategy divides the working interval into multiple equal time slices, so DVRPTD is decomposed into a sequence of static problems. We use an insert heuristic to incorporate new requests into the current solution and a large neighborhood search algorithm (LNS) with various remove and re-insert strategies to optimize at the end of each time slice. The computational results of 22 benchmarks ranging from 50 up to 385 customers prove the superiority of our algorithm comparing with those existing ones.