The multivisit drone routing problem with edge launches: An iterative approach with discrete and continuous improvements

Abstract

In recent years, the usage of drones in last‐mile logistics has stirred great interest in the operations research community. Many papers have considered schemes of hybrid truck‐and‐drone delivery. In this paper, we focus on the Multivisit Drone Routing Problem with Edge Launches (MVDRP‐EL) which assumes: a heterogenous set of packages, a drone capable of carrying multiple packages at a time and that can be launched and retrieved along an edge, a flexible launch/retrieval site set, and a user‐defined energy depletion function. We believe this paper is the first to exploit edge launch ability through a global continuous approach. In this context, we propose an original formulation based on the Covering Salesman Problem to compute a valid lower bound for the problem and an iterative solution method to determine a MVDRP‐EL solution with quality launch/retrieval sites along the road network edges. Each iteration of the solution method consists of two phases. In the first phase, the road network edges are discretized to obtain launch/retrieval sites, and a first solution is determined. In the second phase, the truck route is set and we reduce the completion time by carefully synchronizing truck and drone routes by solving an original Mixed Integer Second Order Cone Program. The lower bound formulation and the proposed method have been tested on several instances and results indicate the effectiveness of the proposed method and the potential value of launching along an edge, respectively.

A new MILP formulation for the flying sidekick traveling salesman problem

“…The research activity on truck-and-drone delivery systems and, in particular, related routing problems have been significant in the last few years. For a complete review on variants of the FS-TSP the interested reader is referred to [16,19], on single-truck multidrone to [6,18], and on multitruck single/multidrone to [24,26]. In this work, we focus our revision on exact solution methods for the FS-TSP.…”

“…Constraints (15) guarantee that an arc coming out of a node i can be traveled by the drone only if the drone serves i or it is the launch node of a sortie. Finally, constraints (16) ensure that an arc entering a node j can be traveled by the drone only if the drone serves j or it is the rendezvous node of a sortie.…”

“…In particular, a bunch of variants of truck-and-drone routing problems has been proposed differing for the side constraints envisaging either the operating conditions or different hybrid truck-and-drone delivery system structures. For the first case, we mention, among the others, the variants involving multiple customers within each drone sortie [20], the dependence of the drone energy consumption on the carried weight [16], and the presence of no-fly zones [27]. For the second case, instead, we cite the variants involving multitruck and multidrone homogeneous and heterogeneous fleets [18,26].…”

Matheuristic for Synchronized Vehicle Routing Problem with Multiple Constraints and Variable Service Time

“…Lastly, the studies of Masone et al (2022) and Gu et al (2022) assume that each drone is capable of visiting multiple customers before returning back to the truck. However, each vehicle is equipped with a single drone.…”

“…Several unresolved realistic challenges in operations research may be solved using VRPMS. Important applications of VRPMS include truck-drone delivery (Masone et al (2022), Meng et al (2023), Chen et al (2021)), healthcare and home care delivery (Kim et al (2017) and Hashemi Doulabi et al (2020)) among others. Solving mathematical models of VRPMS is a challenging task due to the large number of decision variables and constraints involved.…”