Learning Heuristics for Transit Network Design and Improvement with Deep Reinforcement Learning

2024-04-08

arXiv.org

AI Summary

Public transit systems worldwide face dual pressures of budget cuts and declining ridership, necessitating cost-effective network redesigns that preserve service quality. This paper proposes an end-to-end framework integrating deep reinforcement learning (DRL) and graph neural networks (GNNs), the first to automatically learn low-level route modification heuristics—traditionally handcrafted—for evolutionary algorithms. The bus network is modeled as a dynamic graph, with GNNs encoding its topological structure and a DRL agent selecting optimal local modification operations. Evaluated on the Mumford synthetic benchmark, the method achieves new state-of-the-art performance. In large-scale simulations on real-world data from Laval, Canada, it improves passenger accessibility by 52%, on-time performance by 25%, and reduces operational costs by up to 19%. The approach bridges representation learning and combinatorial optimization, enabling data-driven, scalable, and interpretable transit network design.

Technology Category

• Planning, Routing, and Scheduling: Planning with Language Models
• Search and Optimization: Learning to Search
• Application Domains: Transportation

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Economics, Online Markets and Human Computation: Incentives in network design for Web infrastructures and ecosystems
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Transit agencies world-wide face tightening budgets and declining ridership. To maintain quality of service while cutting costs, efficient transit network design is essential. But planning a network of public transit routes is a challenging optimization problem. The most successful approaches to date use metaheuristic algorithms to search through the space of possible transit networks by applying low-level heuristics that randomly alter routes in a network. The design of these low-level heuristics has a major impact on the quality of the result. In this paper we use deep reinforcement learning with graph neural nets to learn low-level heuristics for an evolutionary algorithm, instead of designing them manually. These learned heuristics improve the algorithm's results on benchmark synthetic cities with 70 nodes or more, and achieve new state-of-the-art results the challenging Mumford benchmark. They also improve upon a simulation of the real transit network in the city of Laval, Canada, by as much as 52% and 25% on two key metrics, and offer cost savings of up to 19% over the city's existing transit network.

Problem

• Optimizing transit network design using deep reinforcement learning
• Improving heuristic methods for public transit route planning
• Reducing costs and enhancing efficiency in transit systems

Innovation

• Deep reinforcement learning optimizes transit networks
• Graph neural nets enhance heuristic learning
• Evolutionary algorithm achieves cost-effective solutions