Metaheuristic for Solving the Delivery Man Problem with Drone

keywords: DMPD, metaheuristic, VNS
Delivery Man Problem with Drone (DMPD) is a variant of Delivery Man Problem (DMP). The objective of DMP is to minimize the sum of customers' waiting times. In DMP, there is only a truck to deliver materials to customers while the delivery is completed by collaboration between truck and drone in DMPD. Using a drone is useful when a truck cannot reach some customers in particular circumstances such as narrow roads or natural disasters. For NP-hard problems, metaheuristic is a natural approach to solve medium to large-sized instances. In this paper, a metaheuristic algorithm is proposed. Initially, a solution without drone is created. Then, it is an input of split procedure to convert DMP-solution into DMPD-solution. After that, it is improved by the combination of Variable Neighborhood Search (VNS) and Tabu Search (TS). To explore a new solution space, diversification is applied. The proposed algorithm balances diversification and intensification to prevent the search from local optima. The experimental simulations show that the proposed algorithm reaches good solutions fast, even for large instances.
reference: Vol. 42, 2023, No. 5, pp. 1184–1212