A Matheuristic with Fixed-Sequence Reoptimization for a Real-Life Inventory Routing Problem

Yun He (1) , Christian Artigues (2) , Cyril Briand (2) , Nicolas Jozefowiez (3) , Sandra Ulrich Ngueveu (2)

1 LS2N - Laboratoire des Sciences du Numérique de Nantes
2 LAAS-ROC - Équipe Recherche Opérationnelle, Optimisation Combinatoire et Contraintes
3 LCOMS - Laboratoire de Conception, Optimisation et Modélisation des Systèmes

This paper proposes a matheuristic for solving a real-life Inventory Routing Problem introduced in the ROADEF/EURO Challenge 2016. The method integrates a fixed-sequence mathematical program, two ran-domized greedy algorithms, and a column-generation based heuristic. In particular, the paper discusses the performance of the fixed-sequence mathematical program, which considers a fixed sequence of customer visits and aims at (re)optimizing partial solutions by modifying arrival times and delivered or loaded quantities. Experiments show that the proposed algorithm for the fixed-sequence sub-problem is efficient as a post-optimization process and is even able to improve the best solutions obtained during the Challenge.

Mots clés: Inventory routing problem; Matheuristic; Mixed Integer Linear Fractional Programming

Domaines: Recherche opérationnelle [math.OC]

Yun He, Christian Artigues, Cyril Briand, Nicolas Jozefowiez, Sandra Ulrich Ngueveu. A Matheuristic with Fixed-Sequence Reoptimization for a Real-Life Inventory Routing Problem. Transportation Science, 2020, 54 (2), pp.355-374. ⟨10.1287/trsc.2019.0954⟩. ⟨hal-02944238⟩