A Memetic Algorithm for the Cumulative Capacitated Vehicle Routing Problem Including Priority Indexes

Samuel Nucamendi-Guillén,Diego Flores-Díaz,Elias Olivares-Benitez,Abraham Mendoza

doi:10.3390/app10113943

Samuel Nucamendi-Guillén, Diego Flores-Díaz + Show 2 more

Open Access

https://doi.org/10.3390/app10113943

Copy DOI

Journal: Applied Sciences	Publication Date: Jun 5, 2020
Citations: 14	License type: CC BY 4.0

Affiliation: Universidad Panamericana

Abstract

This paper studies the Cumulative Capacitated Vehicle Routing Problem, including Priority Indexes, a variant of the classical Capacitated Vehicle Routing Problem, which serves the customers according to a certain level of preference. This problem can be effectively implemented in commercial and public environments where customer service is essential, for instance, in the delivery of humanitarian aid or in waste collection systems. For this problem, we aim to minimize two objectives simultaneously, the total latency and the total tardiness of the system. A Mixed Integer formulation is developed and solved using the AUGMECON2 approach to obtain true efficient Pareto fronts. However, as expected, the use of commercial software was able to solve only small instances, up to 15 customers. Therefore, two versions of a Memetic Algorithm with Random Keys (MA-RK) were developed to solve the problem. The computational results show that both algorithms provided good solutions, although the second version obtained denser and higher quality Pareto fronts. Later, both algorithms were used to solve larger instances (20–100 customers). The results were mixed in terms of quality but, in general, the MA-RK v2 consistently outperforms the first version. The models and algorithms proposed in this research provide useful insights for the decision-making process and can be applied to solve a wide variety of business situations where economic, customer service, environmental, and social concerns are involved.

Highlights

In this paper, we study the biobjective variant of the Cumulative Capacitated Vehicle RoutingProblem (CCVRP), a ”customer-centric“ variant of the classical Capacitated Vehicle Routing Problem (CVRP) [1] in which a fleet of k vehicles serves a set of customers by respecting their priority, defined as an index assigned to each node to indicate the preference to be served
The first set of experiments aims at evaluating the performance of the formulation concerning optimality, and the effectiveness of the Memetic Algorithm with Random Keys (MA-RK) comparing the results with those obtained by the resolution of the model
This study addressed the biobjective Cumulative Capacitated Vehicle Routing Problem

Summary

Introduction

Problem (CCVRP), a ”customer-centric“ variant of the classical Capacitated Vehicle Routing Problem (CVRP) [1] in which a fleet of k vehicles serves a set of customers by respecting their priority, defined as an index assigned to each node to indicate the preference to be served. Unlike the traditional VRP, which focuses on the impact that routing costs have on logistics and, in particular, in the transportation activities within the supply chain, the CCVRP rises as a particularization that covers broad objectives centered around service level issues. This problem is relevant in contexts where both customer satisfaction and company profits are prioritized.

Methods

Results

Conclusion