TrafficKAN-GCN: Graph Convolutional-based Kolmogorov-Arnold Network for Traffic Flow Optimization

2025-03-05

AI Summary

To address the performance degradation of conventional models in large-scale dynamic urban traffic networks—caused by high spatiotemporal complexity and strong data noise—this paper proposes the KAN-GCN fusion framework. It pioneers the deep integration of Kolmogorov–Arnold Networks (KANs), which offer adaptive nonlinear fitting, with Graph Convolutional Networks (GCNs), which excel at topological modeling, thereby constructing a robust spatiotemporal graph neural network. The framework effectively handles irregular and noisy traffic flow data, and supports real-time path re-routing and traffic redistribution under sudden disruptions (e.g., bridge collapse). Experiments on real-world traffic data from the Baltimore metropolitan area demonstrate that our method significantly outperforms MLP-GCN, standard GCN, and Transformer-based baselines in prediction accuracy and congestion mitigation. Moreover, it exhibits superior anomaly responsiveness and generalization robustness.

Technology Category

• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems
• Machine Learning: Graph-based Machine Learning
• Data Mining & Knowledge Management: Graph Mining, Social Network Analysis & Community

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Urban traffic optimization is critical for improving transportation efficiency and alleviating congestion, particularly in large-scale dynamic networks. Traditional methods, such as Dijkstra's and Floyd's algorithms, provide effective solutions in static settings, but they struggle with the spatial-temporal complexity of real-world traffic flows. In this work, we propose TrafficKAN-GCN, a hybrid deep learning framework combining Kolmogorov-Arnold Networks (KAN) with Graph Convolutional Networks (GCN), designed to enhance urban traffic flow optimization. By integrating KAN's adaptive nonlinear function approximation with GCN's spatial graph learning capabilities, TrafficKAN-GCN captures both complex traffic patterns and topological dependencies. We evaluate the proposed framework using real-world traffic data from the Baltimore Metropolitan area. Compared with baseline models such as MLP-GCN, standard GCN, and Transformer-based approaches, TrafficKAN-GCN achieves competitive prediction accuracy while demonstrating improved robustness in handling noisy and irregular traffic data. Our experiments further highlight the framework's ability to redistribute traffic flow, mitigate congestion, and adapt to disruptive events, such as the Francis Scott Key Bridge collapse. This study contributes to the growing body of work on hybrid graph learning for intelligent transportation systems, highlighting the potential of combining KAN and GCN for real-time traffic optimization. Future work will focus on reducing computational overhead and integrating Transformer-based temporal modeling for enhanced long-term traffic prediction. The proposed TrafficKAN-GCN framework offers a promising direction for data-driven urban mobility management, balancing predictive accuracy, robustness, and computational efficiency.

Problem

• Optimizing urban traffic flow in dynamic networks
• Handling spatial-temporal complexity in traffic data
• Improving robustness and accuracy in traffic prediction

Innovation

• Combines Kolmogorov-Arnold Networks with Graph Convolutional Networks
• Captures complex traffic patterns and topological dependencies
• Improves robustness in handling noisy, irregular traffic data