Parallel island based Memetic Algorithm with Lin–Kernighan local search for a real-life Two-Echelon Heterogeneous Vehicle Routing Problem based on Brazilian wholesale companies

This paper deals with a Two-Echelon Fixed Fleet Heterogeneous Vehicle Routing Problem (2E-HVRP) faced by Brazilian wholesale companies. Vehicle routing problems with more than one phase are known as Multi-Echelon VRP and consider situations in which freight is moved through some intermediate facilit...

Full description

Saved in:
Bibliographic Details
Published in:Applied soft computing Vol. 76; pp. 697 - 711
Main Authors: Bevilaqua, Andre, Bevilaqua, Diego, Yamanaka, Keiji
Format: Journal Article
Language:English
Published: Elsevier B.V 01.03.2019
Subjects:
Heterogeneous fleet
Island based Memetic Algorithm
Lin–Kernighan
Two-Echelon Vehicle Routing Problem
Wholesale companies
Wholesale companies
Lin–Kernighan
Heterogeneous fleet
Two-Echelon Vehicle Routing Problem
Island based Memetic Algorithm
ISSN:1568-4946, 1872-9681
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:This paper deals with a Two-Echelon Fixed Fleet Heterogeneous Vehicle Routing Problem (2E-HVRP) faced by Brazilian wholesale companies. Vehicle routing problems with more than one phase are known as Multi-Echelon VRP and consider situations in which freight is moved through some intermediate facilities (e.g., cross-docks or distribution centers) before reaching its destination. The first phase of the problem dealt here is to choose a first-level vehicle, from an heterogeneous set, that will leave a depot and reach an intermediate uncapacitated facility (satellite) to serve a set of second-level vehicles. After that, it is necessary to define routes for smaller vehicles, also from an heterogeneous set, that will visit a set of customers departing from and returning to a satellite. The solution proposed here is an efficient island based memetic algorithm with a local search procedure based on Lin–Kernighan heuristic (IBMA-LK). In order to attest the algorithm’s efficiency, first it was tested in single echelon HVRP benchmark instances. After that the instances were adapted for two-echelon context and used for 2E-HVRP validation and, finally, it was tested on 2E-HVRP instances created using real world normalized data. Localsolver tool was also executed for comparison purposes. Promising results (which corroborate results obtained on the real problem) and future works are presented and discussed. •A set of 2E-HVRP instances based on real-world normalized data from wholesale companies.•An adapted object-oriented chromosome representation for the 2E-HVRP context studied.•A modified crossover and an heuristic function that smoothly decreases overload.•An efficient way to use Lin-Kernigham heuristic as an intensive local search strategy.•A demonstration of IBMA-LK competitiveness for HVRP and 2E-HVRP instances.
ISSN:1568-4946
1872-9681
DOI:10.1016/j.asoc.2018.12.036