Towards Fair and Efficient Public Transportation: A Bus Stop Model

2024-11-13

arXiv.org

AI Summary

This paper addresses the bus stop location problem under a fixed number of stops, jointly optimizing system efficiency (total travel time) and group fairness. Passengers may either walk the entire trip or walk to the nearest stop and transfer to bus service; the cost model explicitly incorporates walking sensitivity. Innovatively, we introduce two axiomatic fairness notions from algorithmic fairness—justified representation and approximate core—to bus stop location for the first time, proposing the first joint optimization algorithm that is polynomial-time solvable, efficiency-optimal, and provides provable fairness guarantees. We theoretically establish that the algorithm satisfies strong fairness in walking-dominant scenarios. Empirical evaluations on synthetic and realistic instances demonstrate that the algorithm achieves over 95% coverage of fair solutions while significantly reducing total travel time compared to baseline methods.

Technology Category

• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems
• Search and Optimization: Algorithm Configuration
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization

Application Category

• Responsible Web: Human-perceived consequences of algorithmic deployment on the web
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Social Networks and Social Media: Fairness and bias in social network and social media analysis

Abstract

We consider a stylized formal model of public transportation, where a set of agents need to travel along a given road, and there is a bus that runs the length of this road. Each agent has a left terminal and a right terminal between which they wish to travel; they can walk all the way, or walk to/from the nearest stop and use the bus for the rest of their journey. The bus can make a fixed number of stops, and the planner needs to select locations for these stops. We study notions of efficiency and fairness for this setting. First, we give a polynomial-time algorithm for computing a solution that minimizes the total travel time; our approach can capture further extensions of the base model, such as more general cost functions or existing infrastructure. Second, we develop a polynomial-time algorithm that outputs solutions with provable fairness guarantees (such as a variant of the justified representation axiom or $2$-approximate core) as long as the agents'costs only depend on the distance they need to walk. Our simulations indicate that our algorithm almost always outputs fair solutions, even for parameter regimes that do not admit theoretical guarantees.

Problem

• Optimizing bus stop placement to minimize total travel time
• Developing fair bus stop allocation with representation guarantees
• Balancing efficiency and fairness in public transportation systems

Innovation

• Polynomial-time algorithm minimizes total travel time
• Algorithm provides provable fairness guarantees for agents
• Simulations validate fairness across diverse parameter regimes