Organization of external cares using the Large Neighborhood Search optimization

In this article, we address the organization of external care in multiple sites and, in particular, the planning of the accompaniment of a group of patients with disabilities by a service. To model this VRP-TW, a mixed integer linear programming model is proposed and allows to optimize the travel time of professionals between care sites (schools), taking into account different constraints. These constraints include respect of patient and professional availabilities (time-windows), matching between patient need and professional competencies, capacities of the facilities where the care takes place, as well as scheduling rules specific to this problem. High computation time of the MILP has been observed when the size of the problem increases. In this paper, Large Neighborhood Search is investigated to reduce model computation time and propose an efficient resolution method.