AlphaTransit: Learning to Design City-scale Transit Routes

Hugging Face Daily Papers Papers

Summary

AlphaTransit combines Monte Carlo Tree Search with neural policy-value networks to optimize bus route design by predicting downstream quality without simulator rollouts. It achieves significant service rate improvements on a Bloomington transit benchmark.

Designing a transit network requires many sequential route extension decisions, but their quality is often visible only after the full network is assembled. This delayed-feedback challenge lies at the heart of the Transit Route Network Design Problem (TRNDP), where route interactions can be deceptive: an extension that appears useful locally can create transfer bottlenecks, produce redundant overlap, or reduce overall throughput. To guide route construction under delayed simulator feedback, we introduce AlphaTransit, a search-based planning framework for cityscale bus network design. AlphaTransit couples Monte Carlo Tree Search (MCTS) with a neural policy-value network: the policy proposes route extensions, the value estimates downstream design quality, and search uses these predictions to refine each decision. This provides decision-time lookahead during route construction without running simulator rollouts inside the search tree. We evaluate AlphaTransit on a new Bloomington TRNDP benchmark with realistic road topology and censusderived demand, under mixed and full transit demand settings. In the Bloomington network, AlphaTransit attains the highest service rate in both demand settings, reaching 54.6% and 82.1%, respectively. Relative to reinforcement learning without search, these correspond to 9.9% and 11.4% service rate gains; relative to MCTS without learned guidance, they correspond to 2.5% and 11.2% gains. These results suggest that coupling learned guidance with MCTS is more effective than using either approach alone for transit network design. Our code and data are publicly available in https://github.com/poudel-bibek/AlphaTransit.
Original Article
View Cached Full Text

Cached at: 06/01/26, 07:21 PM

Paper page - AlphaTransit: Learning to Design City-scale Transit Routes

Source: https://huggingface.co/papers/2605.28730

Abstract

AlphaTransit combines Monte Carlo Tree Search with neural policy-value networks to optimize bus route design by predicting downstream quality and enabling lookahead decisions without simulator rollouts.

Designing a transit network requires many sequential route extension decisions, but their quality is often visible only after the full network is assembled. This delayed-feedback challenge lies at the heart of the Transit Route Network Design Problem (TRNDP), where route interactions can be deceptive: an extension that appears useful locally can create transfer bottlenecks, produce redundant overlap, or reduce overall throughput. To guide route construction under delayed simulator feedback, we introduce AlphaTransit, a search-based planning framework for cityscale bus network design. AlphaTransit couplesMonte Carlo Tree Search(MCTS) with aneural policy-value network: the policy proposesroute extensions, the value estimates downstream design quality, and search uses these predictions to refine each decision. This provides decision-time lookahead during route construction without running simulator rollouts inside the search tree. We evaluate AlphaTransit on a new Bloomington TRNDP benchmark with realistic road topology and censusderived demand, under mixed and full transit demand settings. In the Bloomington network, AlphaTransit attains the highestservice ratein both demand settings, reaching 54.6% and 82.1%, respectively. Relative toreinforcement learningwithout search, these correspond to 9.9% and 11.4%service rategains; relative to MCTS without learned guidance, they correspond to 2.5% and 11.2% gains. These results suggest that coupling learned guidance with MCTS is more effective than using either approach alone fortransit network design. Our code and data are publicly available in https://github.com/poudel-bibek/AlphaTransit.

View arXiv pageView PDFProject pageGitHub4Add to collection

Get this paper in your agent:

hf papers read 2605\.28730

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper1

#### matrix-multiply/alphatransit-checkpoints Reinforcement Learning• Updatedabout 5 hours ago

Datasets citing this paper1

#### matrix-multiply/bloomington-tndp Viewer• Updatedabout 5 hours ago • 6.12k

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2605.28730 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles

Smart Transportation Without Neurons -- Fair Metro Network Expansion with Tabular Reinforcement Learning

arXiv cs.LG

Researchers from the University of Amsterdam propose a tabular reinforcement learning approach to the Metro Network Expansion Problem, showing it achieves comparable performance to Deep RL while reducing training episodes by 18x and carbon emissions by 12x on average. The method also incorporates social equity criteria and is evaluated on real-world metro networks in Xi'an and Amsterdam.