VU University
PyVRP High-Performance Vehicle Routing Problem Solver
Pages
15
Time to read
42 mins
Publication
Language
English
Pages
15
Time to read
42 mins
Publication
Language
English
This article is a technical report that introduces PyVRP, a Python package designed for solving vehicle routing problems (VRPs) with a focus on high performance. The package implements a hybrid genetic search algorithm and is particularly tailored for the vehicle routing problem with time windows (VRPTW), while also being extendable to other VRP variants. PyVRP combines the flexibility of Python with the performance of C++ by implementing performance-critical components in C++ and allowing customization at the Python level. The report details the implementation, which has been recognized for its effectiveness, ranking first in notable vehicle routing competitions. The package is structured to be user-friendly and well-documented, enabling researchers and practitioners to easily utilize and adapt the solver for various applications. The article also discusses the software's availability under the MIT license and its potential impact on the field of vehicle routing and machine learning research.