Optimal allocating and sizing of passenger ferry fleet in maritime transport

Abstract

Ferry service represents an important geostrategic, economic and social issues for the coastal regions that are using this mode of transport to satisfy demands of domicile population and tourists. The existing analyses report the significance of ferry service as an alternative mode of transport in narrow places that have a limited road infrastructure. We investigate the organization of transport policy establishing regular passenger ferry fleet service among a given set of routes and predetermined passenger preferences within the defined planning horizon. A Mixed Integer Programming (MIP), more precisely Mixed Integer Linear Programming (MILP) formulation for deterministic optimization problem related to the maximization of ferry operator's profit is developed. It considers the passenger ferry fleet sizing taking into account the employment of the existing and subcontracted fleet. The proposed model can be used to optimally solve instances by using a commercial exact solver. To deal with hard examples, we used three matheuristic methods: Variable Neighborhood Branching (VNB), Variable Neighborhood Decomposition Search for 0-1 MIP (VNDS 0–1 MIP) and Variable Intensity Neighborhood Search (VINS). Our experimental results show that in most of the cases matheuristics produce solutions of equal or higher quality in less computational time compared to the commercial exact solver.