A matheuristic for the multi-product maritime inventory routing problem. (June 2023)

• Record Type: Journal Article
• Authors: Shaabani, Homayoun; Hoff, Arild; Hvattum, Lars Magnus; Laporte, Gilbert
• Is Part Of: Computers & operations research. Volume 154(2023)
• Publication Date: 2023-06
• Subjects: Long planning horizon -- Mixed integer programming -- Maritime inventory routing; Operations research -- Periodicals; Electronic digital computers -- Periodicals; 004.05
• DOI: 10.1016/j.cor.2023.106214
• Languages: English
• ISSNs: 0305-0548

Abstract

Highlights: A two-phase matheuristic for a maritime inventory routing problem is proposed. A multi-product problem for maritime inventory routing is studied. The single-product version of the problem is compared to the multi-product problem. Abstract: This paper considers a multi-product maritime inventory routing problem in which the routes of the vessels are determined while inventories in the ports must remain within given bounds. Since there is more than one product, each port can be a consumption port for one product and a production port for another product. We propose a matheuristic to deal with large-size instances with a planning horizon of 60 days. The first phase of the heuristic identifies a subset of the arcs in order to use them as input parameters for the second phase. In the second phase, a mathematical model is solved with the selected arcs fixed. These two phases are iterated until a stopping criterion is satisfied. The solutions produced by the matheuristic are compared with those obtained by CPLEX. Comparing the single-product version of the problem with the multi-product version revealed that instances with multiple products require significantly more run time compared with the single-product instances.