EURO 2024 Copenhagen
Abstract Submission

EURO-Online login

756. Improving a Hybrid Genetic Search for the Capacitated Vehicle Routing Problem

Invited abstract in session WC-64: Heuristics for Vehicle Routing 1, stream VeRoLog - Vehicle Routing and Logistics.

Wednesday, 12:30-14:00
Room: S16 (building: 101)

Authors (first author is the speaker)

1. Lars Magnus Hvattum
Faculty of Logistics, Molde University College

Abstract

An open-source implementation of a hybrid genetic search is currently among the very best heuristics for the capacitated vehicle routing problem. While being highly efficient and effective, the code is also relatively simple. It is therefore an interesting challenge to improve the performance of the heuristic while making as small changes to the code as possible. Two such improvements can be found by analyzing, respectively, a cross-over operator and a splitting procedure. An improved cross-over operator is shown to reduce gaps to best-known solutions by 4.2%, and an improved splitting procedure is shown to reduce gaps by an additional 3.9%. In both cases, changes to the implementation is in the order of tens of lines of code.

Keywords

Status: accepted

Back to the list of papers