Fairness in Accessibility of Public Service Facilities

Public service facility network design has important implications for the quality and fairness of service provision to different population groups. Racial disparities in service accessibility have been underscored in the political science and geography literature, fostering policy discussions. In this paper, we study a fair stochastic facility location problem with congestion under the max-min principle. We first analyze the price of fairness in a two-location problem and show how it depends on the population group sizes. We then develop a tractable optimization framework for the general multi-location setting. Using demographic data from two major cities in North America, we empirically evaluate the fair solution against the utilitarian solution. Our results show that implementing a fair solution leads to substantial (up to 71%) improvement of fairness measures over the utilitarian solution, with relatively limited impact on the overall service quality