Stork: A Flexible Optimization Tool for Solving Multiple Vehicle Routing Problem Variants

The growing demand for more personalized transportation services has stressed the necessity for flexible and usable Mobility as a Service solutions, especially in urban scenarios. In this context, this paper presents Stork, an optimization-based tool focused on solving, analyzing, and visualizing all relevant extensions of the vehicle routing problem, the formal representation of the dynamics behind many real-life mobility issues. The tool is capable of solving all the well-known variants of the vehicle routing problem, including, among others, capacities, pickup and delivery, heterogeneous feet, time windows, heterogeneous demands, multiple depots, and 3D loading constraints. In addition, it supports any combination of variants, allowing the user to select exactly which ones are needed to represent the problem at hand. Stork also provides many heuristics for generating initial solutions as well as several metaheuristics to improve them, allowing the user to customize the optimization process. This flexibility allows Stork to be able to handle a wide range of situations, from freight transportation to people mobility. As a case study, a police emergency response service in Girona (Spain) is presented, where Stork is used to both generate a set of routes for the patrol cars to follow as well as perform real-time rerouting to respond to emergency situations that arise during the shift. Finally, the main insights and discussion are concisely presented to show the potential of the proposed tool.