Sharing a ride time constraint in a multi-trip dial-a-ride problem. An application to the non-urgent patient transportation problem
Résumé
This article presents a method to find close-to-Pareto-optimal solutions for a multi-objectives round-trip dial-a-ride problem (DARP) decomposed into two dependent sub-problems. In the studied problem, each user expresses two requests, a morning request and an evening request, that represent a round-trip itinerary on one day. The two objectives are the cost and excess ride time minimization. Instead of integrating all the requests into one single DARP, the requests are split between two smaller DARPs: one for the collection of morning requests and one for the collection of evening requests. However, the two DARPs cannot be solved independently because they share a constraint: the user maximal daily ride time. The developed heuristic aims to find a global close to optimal solution within reasonable computation time. The method is applied on real data from the NOMAd project which aims to improve the transportation of disable children from their home to medical centers in the city of Lyon.
Origine : Fichiers produits par l'(les) auteur(s)