A Multi-start Algorithm with Intelligent Neighborhood Selection for solving multi-objective humanitarian vehicle routing problems

In this paper, a problem based on real-world situations in humanitarian logistics is considered, where the main characteristics are the lack of available vehicles and the imperative need of a quick evacuation of all the affected by a disaster, but within the minimum possible travel cost. These aspects will be considered within a Multi-objective Capacitated Vehicle Routing Problem with Multiple Trips, where the objectives under consideration are: minimization of the number of vehicles, of the total travel cost and of the maximum latency. We consider, in this situation, maximum latency to be more relevant than classic latency criteria since reduction of the waiting time of the last affected is crucial for survival when any disaster strikes. For the purpose of producing high-quality solutions, a Multi-start Algorithm with Intelligent Neighborhood Selection is specifically designed and then compared with one of the most competitive reference in the literature, NSGA-II, to prove its superiority.