Dataset for: Dynamic route planning for transportation of reusable goods

Description

This dataset is part of a study submitted to "Computers and Operations Research". The study considers a dynamic vehicle routing problem in which a service provider serves customers that appear over time at different locations in a street network using reusable objects. Vehicles transport these objects between customers, and all routes start and end at a central depot. The objective is to minimize the total distance traveled by all vehicles, the total delivery delay, and the makespan. To solve this problem dynamically, a mixed-integer linear program and an adaptive large neighborhood search (ALNS) heuristic were adapted. Additionally, a novel ALNS variant was proposed. Experiments using instances of different sizes compare the performance of the methods. Furthermore, real-time simulations prove the applicability of the methods in realistic scenarios. The dataset contains two types of instances. Firstly, six instance sets, A to F, with an increasing number of requests, vehicles, and objects, are used to compare the performance of the different solution methods. Each set contains 30 instances. For each instance, a graph is randomly generated by sampling random points in a given rectangle, computing a Delaunay triangulation, and removing some edges at random. By interpreting the edges as line segments, these graphs define a metric space. Both requests and the depot are sampled randomly over both the space defined by the graph and the time. Secondly, one instance set contains 30 instances describing the transport of healthcare equipment between healthcare facilities in Munich. The common graph these instances share is generated as a simplified variant of the street network of Munich in OpenStreetMap. Requests and the depot are sampled only from the positions of healthcare facilities that are hospitals or nursing homes, where each position of such a facility corresponds to the closest point on the street network to the actual location of the facility.

Steps to reproduce

After publication, a link to the paper describing the process of generating this dataset will be provided.

Institutions

• Technical University of Munich

  Bavaria, Munich
• University of Oslo

  Oslo, Oslo

Categories

Funders

• The Research Council of Norway

  Ministry of Education and Research

  Oslo

  Grant ID: 356103

Licence