Routing the Transport of Goods for Institutional Care Customers

Abstract

This thesis focuses on developing an algorithm to solve the routing problem with optional electric vehicles, which takes driver preferences into account for institutional care customer. The heterogeneous fleet, partial charging with heterogeneous charging methods, vehicle capac ity, time windows, and driver scheduling (H-EVRPTW with DS) is introduced as a new variant of a Electric Vehicle Routing Problem. The H-EVRPTWwith DS is formally defined as a Mixed-Integer Lin ear Programming, from which the same formal definition is used for the CP-SAT solver. As previous research indicates that Constraint programming can outperform general Mixed-Integer Linear Program ming solvers. Furthermore, the Simulated Annealing algorithm is pre sented as a meta-heuristic algorithm for the H-EVRPTW with DS. Finally, the hybrid-SA-CP-SAT solver is introduced, which uses the fast Simulated Annealing performance to quickly improve the starting position of the CP-SAT solver. The results show that the CP-SAT solver is not performant enough to solve the H-EVRPTW with DS. Whereas the Simulated Annealing algorithm and the hybrid-SA-CP SAT solver do show promising results.