A Simulated Annealing-based Heuristic for Logistics UAV Scheduling Problem

The UAV scheduling problem is investigated in the paper, in which the packages are assigned to the UAVs and delivered from one distribution station to their destination stations. The optimizing objective is to maximize the satisfaction degrees of the users which are highly related to the packages' deadlines. The problem is applicable in the logistics field. The challenges lie in making decisions on what packages an UAV will deliver in a flight and how many flights will an UAV perform. A flight of an UAV is taken as a flight mission, and the package set delivered in one mission is called a bundle. An iterated heuristic framework is presented to schedule the flight missions which mainly consists of three components: the initial solution generation component (ISG), the first simulated-annealing component (FSA) and the second simulated-annealing component (SSA). ISG generates a feasible solution by the knapsack-based algorithm. FSA and SSA are performed to improve the initial solution on the package level and the bundle level, respectively. By comparing with three related heuristics, we illustrate that the proposal is robust and effective for the problem under study.