Project Details
Projekt
Stochastic Optimization for the Liner Shipping Fleet Repositioning Problem
Subject Area
Accounting and Finance
Term
from 2016 to 2021
Project identifier
Deutsche Forschungsgemeinschaft (DFG) - Project number 346183302
A central problem for liner carriers is the efficient repositioning of container vessels between services in their networks. Finding plans of activities minimizing the cost of repositioning while avoiding the disruption of cargo flows in the network is an NP-hard optimization problem. Recent approaches have solved this problem using deterministic algorithms, however the problem's input data is highly stochastic, involving varying container demands, uncertain travel times, and service disruptions. Solving this problem with stochastic optimization techniques is critical not only for providing more robust solutions to carriers, but also because it represents a significant first step to solving stochastic maritime problems.Many repositionings are carried out each year and each repositioning can cost upwards of a million Euros. Tierney et al. (2014) show potential gains from optimizing a single repositioning plan of up to US$14 million over current liner carrier operations. Furthermore, since one of the main cost components of repositioning ships is bunker fuel, optimizing vessel activities lowers the amount of fuel used, reducing both the CO2 and SOx footprint of liner carriers.The goal of liner shipping repositioning problems is to move vessels from their services in a liner shipping network to a new service such that the fuel costs and port fees are minimized, while cargo intake is maximized and liner shipping specific constraints are respected. These problems are characterized by several maritime specific and problem specific side constraints. First, liner shipping networks consist of regularly scheduled, cyclical routes between ports. These periodic schedules must be enforced when repositioning vessels. Second, vessels may undertake a number of cost saving/revenue earning activities, such as repositioning empty containers, carrying spot cargo or sailing on existing services. Finally, the speed at which each vessels sails must be determined. The fuel consumption of a vessel varies roughly cubically with the speed of the vessel, meaning sailing quickly is significantly more expensive than sailing slowly.This proposal will utilize stochastic optimization techniques to generate repositioning plans that are robust to the many uncertainties in the input data of repositioning problems. This work will be generalizable to other maritime transportation problems with similar structures, and will be the first use of stochastic optimization in the liner shipping literature. We will create exact and heuristic methods for solving these difficult, real-world relevant problems in a robust manner, as well as release the first public dataset with stochastic data for a liner shipping problem. Our results will not only be significant for the research community, but also for the liner shipping industry. Using the algorithms developed in this project, carriers will be able to offer lowerprices to shippers while reducing their environmental impact.
DFG Programme
Research Grants
