Under CopyrightSzal, OliverOliverSzalRubbert, Sebastian Hans PeterSebastian Hans PeterRubbertRizvanolli, AnisaAnisaRizvanolli2025-01-282025-01-282024-09-26https://doi.org/10.24406/publica-4165https://publica.fraunhofer.de/handle/publica/48303310.24406/publica-4165In the paper we examine the performance of a quantum annealer on a single product Maritime Inventory Routing Problem (MIRP) with many-to-many route structure, hard inventory constraints, and end-of-horizon considerations. In today’s competitive market, mathematical optimization is becoming an increasingly popular tool to consider for planning a variety of logistics operations across the industry. MIRPs are a class of mathematical optimization problems with the aim to plan efficient sea trade. The task at hand is to optimize the distribution of a bulk product among supplying and demanding ports with limited inventories by a fleet of heterogeneous vessels. A literature review about problem and modelling variations as well as solution methods is provided. Our considered MIRP variant is well-studied and stands out due to its generality, making it a good candidate to apply and conform to a variety of industry needs. However, one challenge of such models can come by the assumption of a finite planning horizon, penalizing any operations near its end. In order to mitigate such end-of horizon effects and make the model more realistic, we propose a reward for non-empty vessel inventories at the end of the planning horizon. With that we avoid the risk of inducing infeasilities, unlike previously applied methods in the MIRP literature. The main aim of this work is to investigate the capabilities and limitations of Quantum annealing as a new solution method for MIRPs. Previously, this technology has only been considered on simpler problems like the capacitated vehicle routing problem. To this end, we generate a set of test instances, formalized as mixed-integer linear models, and benchmark them on both D-Wave’s quantum annealer and a laptop with CPLEX, while varying the computation time. As one of the main takeaways, we find that on the linear test instances the annealer does not profit from longer computation times.enMaritime Inventory Routing ProblemQuantum AnnealerBenchmarksMaritime LogistikQuantum Computingpresentation