Delay-resistant robust vehicle routing with heterogeneous time windows
Lukas Schürmann,Lukas Metz,Petra Mutzel,Tim Niemann,Sebastian Stiller,Andreas M. Tillmann
DOI: https://doi.org/10.2139/ssrn.4530358
IF: 5.159
2024-01-19
Computers & Operations Research
Abstract:We consider a robust variant of the vehicle routing problem with heterogeneous time windows (RVRP-HTW) with a focus on delay-resistant solutions. Here, customers have different availability time windows for every vehicle and must be provided with a preferably tight appointment window for the planned service. Different vehicles are a possibility to model different days on which one physical vehicle can serve a customer. This is the main reason why different time windows for different vehicles are of high practical relevance. To ensure that the appointment windows are adhered to as much as possible, we introduce a new objective function that penalizes delays. Our novel approach allows us to find solutions that are robust with respect to uncertainties in travel and service times limited by a budget polytope. We present a set-partitioning model, the solution of which is based on column generation and employs a labeling algorithm that integrates robustness into the calculations and is adapted to our problem-specific constraints. In a Monte-Carlo simulation on real-life data, we evaluate this method in terms of runtime and solution quality. Our solutions show very good performance, even if the data is more uncertain than assumed for the optimization, incurring only marginal extra travel time compared to a naive deterministic planning scheme.
computer science, interdisciplinary applications,engineering, industrial,operations research & management science