A public transit network optimization model for equitable access to social services
Description
This repository contains raw data generated for use in Rumpf and Kaul 2021 (referenced below), and includes sets of files for defining transit networks as well as raw data tables generated by the solution algorithm. See the included README for an in-depth explanation of each data set. The main focus of the study was to develop and test a public transit design model for improving equity of access to social services throughout a city. The main case study was based on the Chicago Transit Authority network, with the goal of making minor alterations to the bus fleet assignments in order to improve equity of access to primary health care facilities. A small-scale artificial network was also generated for use in sensitivity analysis. The data sets in this repository include network files used by our hybrid tabu search/simulated annealing solution algorithm in order to solve the social access maximization problem (see the GitHub repository referenced below). Also included are the raw data tables from the CTA and artificial network trial sets. The results of this study indicate that it is indeed possible to significantly increase social service access levels in the least advantaged areas of a community while still guaranteeing that transit service remains near its current level. While improving the access in some areas does require that other areas lose some access, the gains are generally much greater than the losses. Moreover, the losses tend to occur in the areas that already enjoy the greatest levels of access, with the net result being a more even distribution of accessibility levels throughout the city.
Files
Steps to reproduce
The included data sets were generated by the various preprocessing and solution programs which can be found at the GitHub link referenced below. See the included README for an in-depth explanation of the contents of each directory. To summarize, the files in chicago_network.zip define a graph representation of the CTA network. These files were generated based on a combination of census data and data from the Chicago data portal, and were used as inputs for the main solution algorithm. The results of these trials are collected in chicago_results.zip and include tables of the social access maximization objective during each iteration of the search algorithm, as well as the fleet size and community area accessibility metric vectors before and after running the solution algorithm. The files in artificial_network.zip define a small-scale artificial network which was procedurally generated to simulate a small city. This network was used for sensitivity analysis of the social access maximization problem by repeatedly solving the model with varying parameters. The results of these trials are collected in artificial_results.zip.