Implementation of a new matheuristic to solve the vehicle routing problem with simultaneous deliveries and pick-ups – VRPSPD

Abstract

The objective of this article is to present a new methodology for solving the Homogeneous Vehicle Routing Problem with Simultaneous Pickups and Deliveries (VRPSPD). The methodology integrates a matheuristic using the Chu-Beasley genetic algorithm and mixed integer linear programming, based on the Branch-and-Bound procedure. The best configuration obtained from the genetic algorithm is improved through the use of constructive heuristic methods in determining sub-problems, which contribute to the creation of the initial population necessary in the stage of local improvement. The goal is to determine minimum-cost routes that satisfy the pick-up and delivery demands of a group of geographically dispersed customers, while considering the restrictions of the system and the number of vehicles required. The methodology is implemented in C++ and a solver CPLEX software is used to find the solution. Results from test instances in specialized literature demonstrate the efficiency of this new hybrid model, showing good results in short computation times. This article is based on the author's doctoral thesis in engineering and presents similar topics discussed in international congresses, which may explain the similarities in the consulted bibliographical references.