Instances for Time-Dependent Routing​

Welcome to our repository of synthetic and realistic instances for time-dependent routing problems. This resource provides valuable data for researchers and practitioners working on various routing problems that consider the time-dependent nature of travel times. Although Time-Dependent Vehicle Routing problems have received increased attention from the scientific community in recent years, there is still a lack of real time-dependent data. Only big IT players (such as Google, Apple, Microsoft, …) have the availability of high-quality historical time-dependent data. As a result, there are no real travel time dataset freely available to the entire research community. To overcome this aspect, most of the literature on Time-Dependent Vehicle Routing problems relies on synthetic travel time functions. As far as synthetic time-dependent data is concerned, we observe that there are (highly-cited) contributions (see for example Ichoua et al. [2003], Hashimoto et al. [2008]) on vehicle routing problems where the computational campaign relies on time-dependent graphs which satisfy the sufficient conditions partially introduced by Cordeau et al. [2014] and further generalized by the path ranking invariance property defined by Adamo et al. in “On path ranking in time-dependent graphs”. Below, you’ll find descriptions of the available instance groups, along with a detailed PDF explaining the formats used.

PDF Description

We have provided a comprehensive PDF document that outlines the formats of the instances available on this page. This guide will help you understand the structure of the data and how to utilize it effectively in your research or applications.

DOWNLOAD PDF 

Instance Groups

1. Instances from Adamo et al. [2020]:
   • This group includes instances derived from the Time-Dependent Traveling Salesman Problem
   • DOWNLOAD [ sha256 signature 98b1829a628765bee29aa2cdd89e8aaa977eb541861a8ff67a4dbdadf7b10cdf ]
   • Instance list
2. Instances from Arigliano et al. [2019]:
   • This group includes instances derived from the Time-Dependent Traveling Salesman Problem with Time Windows
   • DOWNLOAD  [ sha256 signature bcdaa3454216e2469232c450ee73a65b575be55ee0ec403218026367ae3081db ]
   • Instance list
3. Realistic Instances from Ghiani et al. [2020]:
   • ROME DATASET   [ sha256 signature 4b2de7d535239ed3e70d56ecc9cbd9345d53dbd38407d041394244a4c9a63ed6 ] Instance list
   • PARIS DATASET [ sha256 signature 85401e8e7921039033fbc75c1ad1a4cda6638e153698f7a402918aef03f15614 ] Instance list
   • LONDON DATASET [ sha256 signature 8e77f62466faa2b6a3fd1d51db26f923437f3aef3aafb691a21d6e8a48677b8d ] Instance list
4. Instances from Vu et al. [2020]:
   • This group includes instances derived from the Time-Dependent Traveling Salesman Problem with Time Windows
   • 60 customers  [ sha256 signature 3377511dc9f18cdb5cd1e706966d1fae2f180b7c38300cd1430e3bcf7d59f0a3 ]
   • 80 customers  [ sha256 signature 4b83c6a5a956d85fde58b1c51ac7d94f64074ab366df1b77c6f151a5cff308f6 ]
   • 100 customers [ sha256 signature 97177f52d337610bfc9fb10750f26fa47218e2df4eac66239f11ee027ebaa9ba ]
   • Instance list

We invite you to explore these instances and leverage them in your time-dependent routing research. Download the instances and the accompanying PDF to get started! Additionally, we’ve released an open-source Python library to support research and development in time-dependent routing. You can find it on GitHub. A quick introduction is available in this Colab notebook