Solving MDHFVRP and Developing a User Interface for Agricultural Logistics Planning

Abstract

This study addresses the resolution of a complex Vehicle Routing Problem (VRP) for CPH Agriculture, specifically a Multi-Depot Heterogeneous Capacitated Vehicle Routing Problem (MDHFVRP). The primary objective is to optimize agricultural inputs distribution by minimizing logistics costs while satisfying operational constraints. Two distinct approaches were developed and compared: an exact method using the IBM ILOG OPL CPLEX solver, based on the branch-and-cut algorithm, and an approximate method employing a genetic metaheuristic. The mathematical modeling of the problem was formulated, taking into account the company's specificities such as multiple depots, heterogeneous fleet, and capacity constraints. A comparative analysis of both methods performance was conducted, evaluating their effectiveness in terms of solution quality and computation time for various problem sizes. Results demonstrate that the exact method is more suitable for small-scale problems, offering optimal solutions, while the genetic algorithm proves more efficient for large instances, providing good quality solutions within reasonable computation times. Additionally, a graphical user interface named VRS (Vehicle Routing Solver) was developed to facilitate the solution's use by CPH Agriculture's logisticians