Models and algorithms for reliability-oriented Dial-a-Ride with autonomous electric vehicles

We consider a static decision model related to the management of a Dial-a-Ride (DAR) system involving small autonomous electrical vehicles in a closed industrial site. Because of the specific features of the system, in this paper we concentrate on its reliability and propose a model that aims at assigning requests to vehicles by minimizing the number of loading/unloading operations. We propose an integer linear programming formulation of such Stop Number Minimization Problem and examine the behavior of some of its variants. Next, we consider and analyze a set covering oriented reformulation of the model. Finally, we propose a Greedy Randomized Adaptive Search Procedure (GRASP) based heuristic approach that implements insertion mechanisms and is well fitted to realistic dynamic contexts. All proposed methods are tested on benchmark instances involving some tens of requests.