文章背景与核心概要
城市交通拥堵依然是现代社会面临的重大挑战，它导致生产力下降、出行成本增加以及碳排放上升。尽管全网范围的实时重路由在模拟实验中表现出较好的效果，但它往往建立在一个不切实际的假设之上：即道路上的每一辆车都可以在每个决策时间间隔内被重新规划路线。
为了克服这一局限，本文作者推出了 HLSR，一种新颖的混合实时预测选择性（Hybrid Live-Forecast Selective）车辆重路由框架。HLSR 通过以下方式优化交通流： * 数据融合： 将实时路段车速与短期交通预测相结合。 * 选择性干预： 限制重路由的作用范围以提高工程可行性。 * 先进机制： 利用双阈值拥堵检测、经过校准的上游选择、驾驶员画像定制的行程时间预测，以及基于行程时间加权的 \(k\) 短路径生成算法。
通过采用依赖于预测时界的混合速度指标来进行多成本路径分配，HLSR 为实时拥堵管理提供了一种更具实用性和可扩展性的解决方案。
HLSR: Hybrid Live Forecast Selective Dynamic Vehicle Rerouting for Real-Time Congestion Avoidance
HLSR: Hybrid Live Forecast Selective Dynamic Vehicle Rerouting for Real-Time Congestion Avoidance
Authors: Xiao Wang, Shun Ren Yang, Hui Nien Hung
Date: August 18, 2026
Subject: Artificial Intelligence (cs.AI)
DOI: 10.48550/arXiv.2608.18056
Authors: Xiao Wang, Shun Ren Yang, Hui Nien Hung
Date: August 18, 2026
Subject: Artificial Intelligence (cs.AI)
DOI: 10.48550/arXiv.2608.18056
Summary
Summary
Urban traffic congestion remains a significant challenge, contributing to decreased productivity, increased travel costs, and higher emissions. While network-wide, real-time rerouting is effective in simulations, it often relies on the unrealistic assumption that every vehicle on the road can be replanned at every decision interval.
Urban traffic congestion remains a significant challenge, contributing to decreased productivity, increased travel costs, and higher emissions. While network-wide, real-time rerouting is effective in simulations, it often relies on the unrealistic assumption that every vehicle on the road can be replanned at every decision interval.
The authors introduce HLSR, a novel framework for Hybrid Live-Forecast Selective vehicle rerouting. HLSR optimizes traffic flow by: * Fusing Data: Combining live edge speeds with short-horizon traffic forecasts. * Selective Intervention: Limiting the scope of rerouting to improve feasibility. * Advanced Mechanisms: Utilizing dual-threshold congestion detection, calibrated upstream selection, driver-tailored travel-time prediction, and travel-time-weighted \(k\)-shortest-path generation.
The authors introduce HLSR, a novel framework for Hybrid Live-Forecast Selective vehicle rerouting. HLSR optimizes traffic flow by: * Fusing Data: Combining live edge speeds with short-horizon traffic forecasts. * Selective Intervention: Limiting the scope of rerouting to improve feasibility. * Advanced Mechanisms: Utilizing dual-threshold congestion detection, calibrated upstream selection, driver-tailored travel-time prediction, and travel-time-weighted \(k\)-shortest-path generation.
By employing a horizon-dependent hybrid speed metric for multi-cost route allocation, HLSR provides a more practical and scalable approach to real-time congestion management.
By employing a horizon-dependent hybrid speed metric for multi-cost route allocation, HLSR provides a more practical and scalable approach to real-time congestion management.
Accessing the Paper
Accessing the Paper
Metadata & References
Metadata & References
- Cite as: arXiv:2608.18056 [cs.AI]
- License: Creative Commons Attribution 4.0 International
- Cite as: arXiv:2608.18056 [cs.AI]
- License: Creative Commons Attribution 4.0 International
External Resources
External Resources
Note: This document is based on the submission v1 (18 Aug 2026).
Note: This document is based on the submission v1 (18 Aug 2026).