A Multi-Model Dynamic Selection Framework Using Deep Contextual Bandits for Urban Traffic Flow Prediction in Large-Scale Road Networks

Chen, Silai; Mao, Shengfeng; Zhang, Zongcheng; Zhang, Xiaoyuan; Wu, Yunxia; Jiang, Yangsheng; Yao, Zhihong

doi:10.3390/math14030566

This is an early access version, the complete PDF, HTML, and XML versions will be available soon.

Open AccessArticle

A Multi-Model Dynamic Selection Framework Using Deep Contextual Bandits for Urban Traffic Flow Prediction in Large-Scale Road Networks

by

Silai Chen

¹,

Shengfeng Mao

¹,

Zongcheng Zhang

¹,

Xiaoyuan Zhang

¹,

Yunxia Wu

¹

,

Yangsheng Jiang

^1,2,3

and

Zhihong Yao

^1,2,3,*

¹

School of Transportation and Logistics, Southwest Jiaotong University, Chengdu 611756, China

²

National Engineering Laboratory of Integrated Transportation Big Data Application Technology, Southwest Jiaotong University, Chengdu 611756, China

³

National United Engineering Laboratory of Integrated and Intelligent Transportation, Southwest Jiaotong University, Chengdu 611756, China

^*

Author to whom correspondence should be addressed.

Mathematics 2026, 14(3), 566; https://doi.org/10.3390/math14030566

Submission received: 3 January 2026 / Revised: 25 January 2026 / Accepted: 2 February 2026 / Published: 4 February 2026

(This article belongs to the Special Issue Application of Mathematical Methods to Transportation: Modeling and Analysis)

Download Versions Notes

Abstract

To address the challenge of model selection in large-scale traffic flow prediction tasks, this paper proposes a dynamic multi-model selection framework based on Deep Contextual Bandits (DCB). Centered on the optimal combination of sub-models, the framework leverages contextual information of road segments to select dynamically among candidate predictors, achieving more efficient and accurate traffic flow prediction. Several mechanisms are introduced to improve strategy learning and convergence, including a baseline network, experience replay, double-model estimation, and prioritized experience sampling. A clustering-based strategy is further designed to reduce the search space and enhance the generalization and transferability. Experiments on real-world traffic datasets demonstrate that the proposed framework significantly outperforms traditional static fusion methods, reinforcement learning (RL) baselines, and mainstream spatiotemporal prediction models. In particular, the framework yields a 1.0% improvement in R² and a 3.2% reduction in MAE compared to state-of-the-art baselines, while reducing inference time by 43.1%. Moreover, the proposed framework shows strong capability in adaptive model selection under varying contexts, with ablation studies confirming the effectiveness of its key components.

Keywords: traffic flow prediction; deep contextual bandits; multi-model fusion; dynamic decision-making; reinforcement

Share and Cite

MDPI and ACS Style

Chen, S.; Mao, S.; Zhang, Z.; Zhang, X.; Wu, Y.; Jiang, Y.; Yao, Z. A Multi-Model Dynamic Selection Framework Using Deep Contextual Bandits for Urban Traffic Flow Prediction in Large-Scale Road Networks. Mathematics 2026, 14, 566. https://doi.org/10.3390/math14030566

AMA Style

Chen S, Mao S, Zhang Z, Zhang X, Wu Y, Jiang Y, Yao Z. A Multi-Model Dynamic Selection Framework Using Deep Contextual Bandits for Urban Traffic Flow Prediction in Large-Scale Road Networks. Mathematics. 2026; 14(3):566. https://doi.org/10.3390/math14030566

Chicago/Turabian Style

Chen, Silai, Shengfeng Mao, Zongcheng Zhang, Xiaoyuan Zhang, Yunxia Wu, Yangsheng Jiang, and Zhihong Yao. 2026. "A Multi-Model Dynamic Selection Framework Using Deep Contextual Bandits for Urban Traffic Flow Prediction in Large-Scale Road Networks" Mathematics 14, no. 3: 566. https://doi.org/10.3390/math14030566

APA Style

Chen, S., Mao, S., Zhang, Z., Zhang, X., Wu, Y., Jiang, Y., & Yao, Z. (2026). A Multi-Model Dynamic Selection Framework Using Deep Contextual Bandits for Urban Traffic Flow Prediction in Large-Scale Road Networks. Mathematics, 14(3), 566. https://doi.org/10.3390/math14030566

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Menu

A Multi-Model Dynamic Selection Framework Using Deep Contextual Bandits for Urban Traffic Flow Prediction in Large-Scale Road Networks

Abstract

Share and Cite

Article Metrics

Article Access Statistics

Further Information

Guidelines

MDPI Initiatives

Follow MDPI