Operations Research 2025
Abstract Submission

231. A Distributionally Robust Optimization Approach for Liner Shipping Fleet Repositioning

Invited abstract in session WC-1: Port and shipping logistics, stream Mobility, Transportation, and Traffic.

Wednesday, 13:30-15:00
Room: Audimax

Authors (first author is the speaker)

1. Jana Ksciuk
Europa-Universität Viadrina
2. Achim Koberstein
Information and Operations Management, European University Viadrina Frankfurt (Oder)

Abstract

Liner shipping repositioning is the costly process of moving container ships between services within a liner shipping network to adapt the network to fluctuating customer demands. Existing deterministic models for the liner shipping fleet repositioning problem (LSFRP) neglect the inherent uncertainty present in the input parameters. This paper introduces an optimization model for the stochastic LSFRP that addresses uncertainty concerning container demands and ship travel times. Two prevalent paradigms for tackling problems affected by uncertainty are classical stochastic programming (SP) and robust optimization (RO). However, SP often faces challenges with large-scale decision problems and the reliable estimation of the underlying probability distribution. In contrast, RO focuses on the worst-case scenario and is often considered overly conservative.
We employ an alternative modeling paradigm known as distributionally robust optimization (DRO), which combines the strengths of SP and RO. DRO mitigates the "optimizer's curse" characteristic of stochastic programming by adopting a worst-case approach. Furthermore, distributionally robust models are often tractable and can be solved efficiently. We propose a two-stage distributionally robust optimization approach where we utilize the Wasserstein metric to construct an ambiguity set of possible distributions based on a single empirical distribution. The objective of the distributionally robust LSFRP is to minimize the worst-case expected cost with respect to this ambiguity set.

Keywords

Status: accepted

Back to the list of papers