Transportation Research Part C-Emerging Technologies最新文献_第5页

U-Aerodrome: Data-driven and risk-bounded airspace reconfiguration for safe integration of urban air mobility at aerodrome U-Aerodrome：数据驱动和风险有限的空域重构，用于机场城市空中机动性的安全集成

IF 7.6 1区工程技术 Q1 TRANSPORTATION SCIENCE & TECHNOLOGY

Transportation Research Part C-Emerging Technologies

Pub Date : 2026-01-07 DOI: 10.1016/j.trc.2025.105506

Xinting Hu , Bizhao Pang , Sameer Alam , Mir Feroskhan

Urban Air Mobility (UAM) offers promising solutions for alleviating urban congestion and enabling seamless air transportation. However, its integration near aerodromes is limited by static no-fly zones and traditional airspace management practices. Existing boundary-setting methods often depend on oversimplified assumptions about trajectory distributions or apply rigid spatial constraints, which can lead to safety risks and inefficient airspace utilization. To address these limitations, this study introduces U-Aerodrome, a data-driven and risk-bounded airspace reconfiguration framework designed to support the safe and flexible integration of UAM operations near controlled aerodromes. The approach employs procedure-based trajectory classification and equal-altitude sampling to ensure equitable and non-biased representation of flight patterns. It further incorporates probabilistic boundary estimation that accommodates both Gaussian and non-Gaussian distributions, as well as a time-dependent boundary update mechanism responsive to dynamic traffic demand. The framework is validated using real-world data collected from Singapore Changi Airport. Results show that U-Aerodrome reduces missed detections and conservative volume compared to a purely Gaussian baseline, yielding 30.95 % average safety improvement and 15.25 % higher availability. The time-dependent mechanism further reduces unnecessary restrictions by an additional 20.02 % on average compared with baselines assuming static boundaries. The framework supports flexible and statistically grounded planning for safe UAM access near aerodromes.

城市空中交通（UAM）为缓解城市拥堵和实现无缝空中运输提供了有前途的解决方案。然而，它在机场附近的整合受到静态禁飞区和传统空域管理实践的限制。现有的边界设置方法往往依赖于对轨迹分布的过于简化的假设或应用刚性的空间约束，这可能导致安全风险和低效的空域利用。为了解决这些限制，本研究引入了U-Aerodrome，这是一个数据驱动和风险有限的空域重构框架，旨在支持受控机场附近UAM操作的安全和灵活集成。该方法采用基于程序的轨迹分类和等高度抽样，以确保公平和无偏见地表示飞行模式。它进一步结合了适应高斯和非高斯分布的概率边界估计，以及响应动态交通需求的时变边界更新机制。该框架使用从新加坡樟宜机场收集的真实数据进行验证。结果表明，与纯高斯基线相比，U-Aerodrome减少了漏检和保守体积，平均安全性提高了30.95%，可用性提高了15.25%。与假设静态边界的基线相比，依赖时间的机制进一步减少了不必要的限制，平均减少了20.02%。该框架支持机场附近安全UAM通道的灵活和统计基础规划。

{"title":"U-Aerodrome: Data-driven and risk-bounded airspace reconfiguration for safe integration of urban air mobility at aerodrome","authors":"Xinting Hu , Bizhao Pang , Sameer Alam , Mir Feroskhan","doi":"10.1016/j.trc.2025.105506","DOIUrl":"10.1016/j.trc.2025.105506","url":null,"abstract":"<div><div>Urban Air Mobility (UAM) offers promising solutions for alleviating urban congestion and enabling seamless air transportation. However, its integration near aerodromes is limited by static no-fly zones and traditional airspace management practices. Existing boundary-setting methods often depend on oversimplified assumptions about trajectory distributions or apply rigid spatial constraints, which can lead to safety risks and inefficient airspace utilization. To address these limitations, this study introduces U-Aerodrome, a data-driven and risk-bounded airspace reconfiguration framework designed to support the safe and flexible integration of UAM operations near controlled aerodromes. The approach employs procedure-based trajectory classification and equal-altitude sampling to ensure equitable and non-biased representation of flight patterns. It further incorporates probabilistic boundary estimation that accommodates both Gaussian and non-Gaussian distributions, as well as a time-dependent boundary update mechanism responsive to dynamic traffic demand. The framework is validated using real-world data collected from Singapore Changi Airport. Results show that U-Aerodrome reduces missed detections and conservative volume compared to a purely Gaussian baseline, yielding 30.95 % average safety improvement and 15.25 % higher availability. The time-dependent mechanism further reduces unnecessary restrictions by an additional 20.02 % on average compared with baselines assuming static boundaries. The framework supports flexible and statistically grounded planning for safe UAM access near aerodromes.</div></div>","PeriodicalId":54417,"journal":{"name":"Transportation Research Part C-Emerging Technologies","volume":"184 ","pages":"Article 105506"},"PeriodicalIF":7.6,"publicationDate":"2026-01-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145927844","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

CONTINA: Confidence interval for traffic demand prediction with coverage guarantee CONTINA：具有覆盖保证的交通需求预测的置信区间

IF 7.6 1区工程技术 Q1 TRANSPORTATION SCIENCE & TECHNOLOGY

Transportation Research Part C-Emerging Technologies

Pub Date : 2025-12-30 DOI: 10.1016/j.trc.2025.105502

Chao Yang , Xiannan Huang , Shuhan Qiu , Yan Cheng

Accurate short-term traffic demand prediction is critical for the operation of traffic systems. Besides point estimation, the confidence interval of the prediction is also of great importance. Many models for traffic operations, such as shared bike rebalancing and taxi dispatching, take into account the uncertainty of future demand and require confidence intervals as the input. However, existing methods for confidence interval modeling rely on strict assumptions, such as unchanging traffic patterns and correct model specifications, to guarantee enough coverage. Therefore, the confidence intervals provided could be invalid, especially in a changing traffic environment. To fill this gap, we propose an efficient method, CONTINA (Conformal Traffic Intervals with Adaptation) to provide interval predictions that can adapt to external changes. By collecting the errors of interval during deployment, the method can adjust the interval in the next step by widening it if the errors are too large or shortening it otherwise. Furthermore, we theoretically prove that the coverage of the confidence intervals provided by our method converges to the target coverage level. Experiments across four real-world datasets and prediction models demonstrate that the proposed method can provide valid confidence intervals with shorter lengths. Our method can help traffic management personnel develop a more reasonable and robust operation plan in practice. And we release the code, model and dataset in GitHub.

准确的短期交通需求预测对交通系统的运行至关重要。除了点估计外，预测的置信区间也很重要。许多交通运营模型，如共享单车再平衡和出租车调度，都考虑了未来需求的不确定性，并需要置信区间作为输入。然而，现有的置信区间建模方法依赖于严格的假设，如不变的交通模式和正确的模型规范，以保证足够的覆盖。因此，所提供的置信区间可能是无效的，特别是在不断变化的流量环境中。为了填补这一空白，我们提出了一种有效的方法CONTINA （Conformal Traffic Intervals with Adaptation）来提供能够适应外部变化的区间预测。该方法通过收集部署时的间隔误差，如果误差太大，则可以扩大间隔，否则可以缩短间隔，从而在下一步调整间隔。进一步，从理论上证明了该方法提供的置信区间的覆盖收敛于目标覆盖水平。在四个实际数据集和预测模型上的实验表明，该方法可以提供较短长度的有效置信区间。该方法可以帮助交通管理人员在实践中制定更合理、更稳健的运营计划。我们在GitHub上发布代码、模型和数据集。

{"title":"CONTINA: Confidence interval for traffic demand prediction with coverage guarantee","authors":"Chao Yang , Xiannan Huang , Shuhan Qiu , Yan Cheng","doi":"10.1016/j.trc.2025.105502","DOIUrl":"10.1016/j.trc.2025.105502","url":null,"abstract":"<div><div>Accurate short-term traffic demand prediction is critical for the operation of traffic systems. Besides point estimation, the confidence interval of the prediction is also of great importance. Many models for traffic operations, such as shared bike rebalancing and taxi dispatching, take into account the uncertainty of future demand and require confidence intervals as the input. However, existing methods for confidence interval modeling rely on strict assumptions, such as unchanging traffic patterns and correct model specifications, to guarantee enough coverage. Therefore, the confidence intervals provided could be invalid, especially in a changing traffic environment. To fill this gap, we propose an efficient method, CONTINA (Conformal Traffic Intervals with Adaptation) to provide interval predictions that can adapt to external changes. By collecting the errors of interval during deployment, the method can adjust the interval in the next step by widening it if the errors are too large or shortening it otherwise. Furthermore, we theoretically prove that the coverage of the confidence intervals provided by our method converges to the target coverage level. Experiments across four real-world datasets and prediction models demonstrate that the proposed method can provide valid confidence intervals with shorter lengths. Our method can help traffic management personnel develop a more reasonable and robust operation plan in practice. And we release the code, model and dataset in <span><span>GitHub</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":54417,"journal":{"name":"Transportation Research Part C-Emerging Technologies","volume":"184 ","pages":"Article 105502"},"PeriodicalIF":7.6,"publicationDate":"2025-12-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145886225","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Uncertainty quantification for joint demand prediction of multi-modal ride-sourcing services using spatiotemporal Mixture-of-Expert neural network 基于时空混合专家神经网络的多模式拼车服务联合需求预测的不确定性量化

IF 7.6 1区工程技术 Q1 TRANSPORTATION SCIENCE & TECHNOLOGY

Transportation Research Part C-Emerging Technologies

Pub Date : 2025-12-29 DOI: 10.1016/j.trc.2025.105507

Xiaobing Liu , Yu Duan , Yangli-ao Geng , Yun Wang , Qingyong LI , Xuedong Yan , Ziyou Gao

Given the sparse uncertainty and highly imbalanced distribution of origin-destination demand in ride-sourcing, probabilistic prediction methods offer rich information to enhance operation decision-making reliability. However, existing uncertainty quantification methods face critical limitations: parametric approaches struggle with irregularities of real-world demand distributions. In contrast, nonparametric methods based on quantile regression are scarcely implemented and suffer from quantile crossing, compromising probabilistic consistency. Furthermore, current approaches seldom consider the crucial inter-service correlations among ride-sourcing services. This paper introduces the Spatiotemporal Mixture-of-Experts Residual-based Multi-quantile Regressive Network (ST-MoE-RMQRN).¹ This novel hybrid framework integrates distribution-free quantile estimation with service correlation modeling. Our three main contributions are: 1) a Residual-based Multi-quantile Regressor that employs monotonicity-constrained residual fitting to ensure non-crossing quantiles while preserving distributional flexibility; 2) an environment-aware MoE architecture that captures service correlations through multiple shared-expert networks, addressing potential demand switches among service modals with sensitivity to temporal heterogeneity and exogenous contextual factors such as weather conditions; 3) comprehensive validation on two real-world ride-sourcing datasets demonstrating that our approach consistently outperforms recent probabilistic predictive models in both deterministic and probabilistic forecast metrics, including the Spatiotemporal Zero-Inflated Negative Binomial Graph Neural Network (STZINB-GNN). Additionally, data-wise and component-wise ablation studies confirm the effectiveness of the entire framework and each part of it, showing that explicitly modeling inter-service dependencies substantially improves both deterministic accuracy and probabilistic calibration. Our model is highly efficient in representing extreme demand fluctuations and data-sparse scenarios, which promise to address existing operational inefficiencies faced by transitioning Transportation Network Companies from reactive to anticipatory decision-making.

考虑到网约车需求的稀疏不确定性和高度不均衡分布，概率预测方法提供了丰富的信息，提高了运营决策的可靠性。然而，现有的不确定性量化方法面临着严重的局限性：参数化方法与现实世界需求分布的不规则性作斗争。相比之下，基于分位数回归的非参数方法很少实现，并且存在分位数交叉，影响概率一致性。此外，目前的方法很少考虑乘车外包服务之间至关重要的服务间相关性。本文介绍了基于时空专家混合残差的多分位数回归网络（ST-MoE-RMQRN）该框架将无分布分位数估计与服务相关建模相结合。我们的三个主要贡献是：1)基于残差的多分位数回归器，它采用单调性约束残差拟合来确保不交叉分位数，同时保持分布的灵活性；2)环境感知的MoE架构，通过多个共享专家网络捕获服务相关性，解决服务模式之间潜在的需求转换，对时间异质性和外生上下文因素（如天气条件）敏感；3)在两个现实世界的打车数据集上进行全面验证，表明我们的方法在确定性和概率预测指标上始终优于最近的概率预测模型，包括时空零膨胀负二项图神经网络（STZINB-GNN）。此外，数据和组件的消融研究证实了整个框架及其每个部分的有效性，表明显式建模服务间依赖关系大大提高了确定性精度和概率校准。我们的模型在表示极端需求波动和数据稀疏场景方面非常高效，有望解决交通网络公司从被动决策向预期决策转变所面临的现有运营效率低下问题。

{"title":"Uncertainty quantification for joint demand prediction of multi-modal ride-sourcing services using spatiotemporal Mixture-of-Expert neural network","authors":"Xiaobing Liu , Yu Duan , Yangli-ao Geng , Yun Wang , Qingyong LI , Xuedong Yan , Ziyou Gao","doi":"10.1016/j.trc.2025.105507","DOIUrl":"10.1016/j.trc.2025.105507","url":null,"abstract":"<div><div>Given the sparse uncertainty and highly imbalanced distribution of origin-destination demand in ride-sourcing, probabilistic prediction methods offer rich information to enhance operation decision-making reliability. However, existing uncertainty quantification methods face critical limitations: parametric approaches struggle with irregularities of real-world demand distributions. In contrast, nonparametric methods based on quantile regression are scarcely implemented and suffer from quantile crossing, compromising probabilistic consistency. Furthermore, current approaches seldom consider the crucial inter-service correlations among ride-sourcing services. This paper introduces the Spatiotemporal Mixture-of-Experts Residual-based Multi-quantile Regressive Network (ST-MoE-RMQRN).<span><span><sup>1</sup></span></span> This novel hybrid framework integrates distribution-free quantile estimation with service correlation modeling. Our three main contributions are: 1) a Residual-based Multi-quantile Regressor that employs monotonicity-constrained residual fitting to ensure non-crossing quantiles while preserving distributional flexibility; 2) an environment-aware MoE architecture that captures service correlations through multiple shared-expert networks, addressing potential demand switches among service modals with sensitivity to temporal heterogeneity and exogenous contextual factors such as weather conditions; 3) comprehensive validation on two real-world ride-sourcing datasets demonstrating that our approach consistently outperforms recent probabilistic predictive models in both deterministic and probabilistic forecast metrics, including the Spatiotemporal Zero-Inflated Negative Binomial Graph Neural Network (STZINB-GNN). Additionally, data-wise and component-wise ablation studies confirm the effectiveness of the entire framework and each part of it, showing that explicitly modeling inter-service dependencies substantially improves both deterministic accuracy and probabilistic calibration. Our model is highly efficient in representing extreme demand fluctuations and data-sparse scenarios, which promise to address existing operational inefficiencies faced by transitioning Transportation Network Companies from reactive to anticipatory decision-making.</div></div>","PeriodicalId":54417,"journal":{"name":"Transportation Research Part C-Emerging Technologies","volume":"184 ","pages":"Article 105507"},"PeriodicalIF":7.6,"publicationDate":"2025-12-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145885537","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Overcoming computational challenges in air transportation: A quantum computing perspective of the status quo and future applicability 克服航空运输中的计算挑战：现状和未来适用性的量子计算视角

IF 7.6 1区工程技术 Q1 TRANSPORTATION SCIENCE & TECHNOLOGY

Transportation Research Part C-Emerging Technologies

Pub Date : 2025-12-27 DOI: 10.1016/j.trc.2025.105505

Zhuoming Du, Sebastian Wandelt, Xiaoqian Sun

Recent research breakthroughs in quantum computing, such as Microsoft’s topological qubits, hold the promise of revolutionizing complex optimization problems, particularly in the air transportation industry. This study aims to estimate the mid-term scalability of quantum computing in air transportation, focusing on prevalent optimization problems including network design, airline scheduling, and gate assignment. These problems are computationally intensive and often intractable for classical computers due to their highly combinatorial nature. We develop a framework to assess the potential scalability of quantum algorithms for these problems, considering factors such as qubit count and error rates. Our findings suggest that significant advancements in quantum hardware and algorithms are necessary before quantum computing can outperform classical methods in this domain. Therefore, while quantum computing offers a promising tool for solving complex optimization problems in air transportation, its real-world application remains a distant goal. We believe that our work helps guiding researchers and industry professionals in their pursuit of quantum-enhanced air transport solutions.

最近在量子计算方面的研究突破，如微软的拓扑量子比特，有望彻底改变复杂的优化问题，特别是在航空运输行业。本研究旨在评估量子计算在航空运输中的中期可扩展性，重点关注网络设计、航班调度和登机口分配等常见优化问题。这些问题是计算密集型的，由于它们的高度组合性，对于经典计算机来说往往是难以处理的。我们开发了一个框架来评估这些问题的量子算法的潜在可扩展性，考虑到诸如量子比特计数和错误率等因素。我们的研究结果表明，在量子计算能够超越该领域的经典方法之前，量子硬件和算法的重大进步是必要的。因此，虽然量子计算为解决航空运输中复杂的优化问题提供了一个有前途的工具，但其在现实世界中的应用仍然是一个遥远的目标。我们相信，我们的工作有助于指导研究人员和行业专业人士寻求量子增强航空运输解决方案。

{"title":"Overcoming computational challenges in air transportation: A quantum computing perspective of the status quo and future applicability","authors":"Zhuoming Du, Sebastian Wandelt, Xiaoqian Sun","doi":"10.1016/j.trc.2025.105505","DOIUrl":"10.1016/j.trc.2025.105505","url":null,"abstract":"<div><div>Recent research breakthroughs in quantum computing, such as Microsoft’s topological qubits, hold the promise of revolutionizing complex optimization problems, particularly in the air transportation industry. This study aims to estimate the mid-term scalability of quantum computing in air transportation, focusing on prevalent optimization problems including network design, airline scheduling, and gate assignment. These problems are computationally intensive and often intractable for classical computers due to their highly combinatorial nature. We develop a framework to assess the potential scalability of quantum algorithms for these problems, considering factors such as qubit count and error rates. Our findings suggest that significant advancements in quantum hardware and algorithms are necessary before quantum computing can outperform classical methods in this domain. Therefore, while quantum computing offers a promising tool for solving complex optimization problems in air transportation, its real-world application remains a distant goal. We believe that our work helps guiding researchers and industry professionals in their pursuit of quantum-enhanced air transport solutions.</div></div>","PeriodicalId":54417,"journal":{"name":"Transportation Research Part C-Emerging Technologies","volume":"184 ","pages":"Article 105505"},"PeriodicalIF":7.6,"publicationDate":"2025-12-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145842860","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Robotic sorting systems: Robot management and layout design optimization 机器人分拣系统：机器人管理和布局设计优化

IF 7.6 1区工程技术 Q1 TRANSPORTATION SCIENCE & TECHNOLOGY

Transportation Research Part C-Emerging Technologies

Pub Date : 2025-12-26 DOI: 10.1016/j.trc.2025.105500

Tong Zhao , Xi Lin , Fang He , Hanwen Dai

In the contemporary logistics industry, automation plays a pivotal role in enhancing production efficiency and expanding industrial scale. In particular, autonomous mobile robots have become integral to modernization efforts in warehouses. One noteworthy application in robotic warehousing is the robotic sorting system (RSS), which is distinguished by its cost-effectiveness, simplicity, scalability, and adaptable throughput control. Previous research on RSS efficiency often assumed an ideal robot management system, ignoring potential traffic delays and assuming constant travel times. To address this gap, we introduce a novel robot traffic management method, named Rhythmic Control for the Sorting Scenario (RC-S), for RSS operations, along with an analytical estimation formula that establishes the quantitative relationship between system performance and configurations. Simulations validate that RC-S reduces average service time by 10.3 % compared to the classical cooperative A* algorithm, while also improving throughput and runtime. Based on the performance analysis of RC-S, we develop a layout optimization model that considers system configurations, desired throughput, and costs to minimize expenses and determine the optimal layout. Numerical studies show that facility costs dominate at lower throughput levels, while labor costs prevail at higher throughput levels. Additionally, due to traffic efficiency limitations, an RSS is well-suited for small-scale operations like end-of-supply-chain distribution centers.

在现代物流业中，自动化对提高生产效率和扩大产业规模起着举足轻重的作用。特别是，自主移动机器人已经成为仓库现代化工作不可或缺的一部分。机器人仓储中一个值得注意的应用是机器人分拣系统（RSS），其特点是成本效益高、简单、可扩展和适应性强的吞吐量控制。以往的RSS效率研究通常假设一个理想的机器人管理系统，忽略潜在的交通延迟，并假设恒定的旅行时间。为了解决这一差距，我们引入了一种新的机器人交通管理方法，称为排序场景的节奏控制（RC-S），用于RSS操作，以及一个分析估计公式，该公式建立了系统性能和配置之间的定量关系。仿真结果表明，RC-S算法与传统的协同A*算法相比，平均服务时间减少了10.3%，同时提高了吞吐量和运行时间。基于RC-S的性能分析，我们建立了考虑系统配置、期望吞吐量和成本的布局优化模型，以最小化费用并确定最佳布局。数值研究表明，设备成本在低吞吐量水平下占主导地位，而劳动力成本在高吞吐量水平下占主导地位。此外，由于交通效率的限制，RSS非常适合像供应链末端配送中心这样的小规模操作。

{"title":"Robotic sorting systems: Robot management and layout design optimization","authors":"Tong Zhao , Xi Lin , Fang He , Hanwen Dai","doi":"10.1016/j.trc.2025.105500","DOIUrl":"10.1016/j.trc.2025.105500","url":null,"abstract":"<div><div>In the contemporary logistics industry, automation plays a pivotal role in enhancing production efficiency and expanding industrial scale. In particular, autonomous mobile robots have become integral to modernization efforts in warehouses. One noteworthy application in robotic warehousing is the robotic sorting system (RSS), which is distinguished by its cost-effectiveness, simplicity, scalability, and adaptable throughput control. Previous research on RSS efficiency often assumed an ideal robot management system, ignoring potential traffic delays and assuming constant travel times. To address this gap, we introduce a novel robot traffic management method, named Rhythmic Control for the Sorting Scenario (RC-S), for RSS operations, along with an analytical estimation formula that establishes the quantitative relationship between system performance and configurations. Simulations validate that RC-S reduces average service time by 10.3 % compared to the classical cooperative A* algorithm, while also improving throughput and runtime. Based on the performance analysis of RC-S, we develop a layout optimization model that considers system configurations, desired throughput, and costs to minimize expenses and determine the optimal layout. Numerical studies show that facility costs dominate at lower throughput levels, while labor costs prevail at higher throughput levels. Additionally, due to traffic efficiency limitations, an RSS is well-suited for small-scale operations like end-of-supply-chain distribution centers.</div></div>","PeriodicalId":54417,"journal":{"name":"Transportation Research Part C-Emerging Technologies","volume":"184 ","pages":"Article 105500"},"PeriodicalIF":7.6,"publicationDate":"2025-12-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145842859","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Integrating first-and-last-mile feeder services with urban public transportation considering transfer time uncertainty 考虑换乘时间不确定性，将首末英里支线服务与城市公共交通相结合

IF 7.6 1区工程技术 Q1 TRANSPORTATION SCIENCE & TECHNOLOGY

Transportation Research Part C-Emerging Technologies

Pub Date : 2025-12-26 DOI: 10.1016/j.trc.2025.105499

Bo Sun, Qiang Meng

This study proposes an integrated first-and-last-mile feeder service (IFLMFS) that coordinates shared mobility solutions with urban public transportation (UPT). Riders can request travel services in the form of a single first-mile (FM) trip prior to UPT, or a single last-mile (LM) trip following UPT usage, or a combination of both, to connect with UPT hubs seamlessly. We formulate the IFLMFS problem as a tailored path-based set-packing (PSC) model. For riders requiring both FM and LM connections, new coupling constraints are introduced into the PSC model to jointly optimize routing decisions, which ensures the chronological consistency between FM and LM routes. To account for the transfer time uncertainty during UPT, arising from factors such as riders’ walking speed variations and boarding behaviors, an effective robust optimization approach is adopted and the PSC model is reformulated with tractable robust counterparts. To efficiently process real-time travel requests, we develop a Column Generation method embedded within a Rolling Horizon Framework (CG-RHF). To accelerate CG-RHF, a time-based pruning strategy is created to tighten the solution space and a decomposition strategy for pricing subproblems is designed across different UPT hubs in parallel. To validate the computational performance of the proposed CG-RHF, we benchmark it against an exact branch-price-and-cut algorithm and a customized adaptive large neighborhood search metaheuristic. Extensive numerical experiments from Singapore’s metro network demonstrate that the integrated service can efficiently deliver high-quality solutions across a variety of rider travel scenarios.

本研究提出了一种集成的第一和最后一英里支线服务（IFLMFS），该服务将共享移动解决方案与城市公共交通（UPT）相协调。乘客可以在UPT之前要求一次第一英里（FM）旅行，或者在UPT使用之后要求一次最后一英里（LM）旅行，或者两者的结合，以无缝连接UPT枢纽。我们将IFLMFS问题表述为一个定制的基于路径的集包装（PSC）模型。对于同时需要FM和LM连接的乘客，在PSC模型中引入新的耦合约束，共同优化路线决策，保证FM和LM路线在时间上的一致性。为了考虑UPT期间乘客步行速度变化和乘车行为等因素引起的换乘时间不确定性，采用了一种有效的鲁棒优化方法，将PSC模型重构为可处理的鲁棒模型。为了有效地处理实时旅行请求，我们开发了一种嵌入滚动地平线框架（CG-RHF）的列生成方法。为了加速CG-RHF，创建了基于时间的剪枝策略来压缩解空间，并在不同的UPT集线器上并行设计了定价子问题的分解策略。为了验证所提出的CG-RHF的计算性能，我们将其与精确的分支价格切割算法和自定义自适应大邻域搜索元启发式算法进行了基准测试。新加坡地铁网络的大量数值实验表明，综合服务可以有效地为各种乘客出行场景提供高质量的解决方案。

{"title":"Integrating first-and-last-mile feeder services with urban public transportation considering transfer time uncertainty","authors":"Bo Sun, Qiang Meng","doi":"10.1016/j.trc.2025.105499","DOIUrl":"10.1016/j.trc.2025.105499","url":null,"abstract":"<div><div>This study proposes an integrated first-and-last-mile feeder service (IFLMFS) that coordinates shared mobility solutions with urban public transportation (UPT). Riders can request travel services in the form of a single first-mile (FM) trip prior to UPT, or a single last-mile (LM) trip following UPT usage, or a combination of both, to connect with UPT hubs seamlessly. We formulate the IFLMFS problem as a tailored path-based set-packing (PSC) model. For riders requiring both FM and LM connections, new coupling constraints are introduced into the PSC model to jointly optimize routing decisions, which ensures the chronological consistency between FM and LM routes. To account for the transfer time uncertainty during UPT, arising from factors such as riders’ walking speed variations and boarding behaviors, an effective robust optimization approach is adopted and the PSC model is reformulated with tractable robust counterparts. To efficiently process real-time travel requests, we develop a Column Generation method embedded within a Rolling Horizon Framework (CG-RHF). To accelerate CG-RHF, a time-based pruning strategy is created to tighten the solution space and a decomposition strategy for pricing subproblems is designed across different UPT hubs in parallel. To validate the computational performance of the proposed CG-RHF, we benchmark it against an exact branch-price-and-cut algorithm and a customized adaptive large neighborhood search metaheuristic. Extensive numerical experiments from Singapore’s metro network demonstrate that the integrated service can efficiently deliver high-quality solutions across a variety of rider travel scenarios.</div></div>","PeriodicalId":54417,"journal":{"name":"Transportation Research Part C-Emerging Technologies","volume":"184 ","pages":"Article 105499"},"PeriodicalIF":7.6,"publicationDate":"2025-12-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145824270","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Dynamic route redundancy-oriented strategic planning towards resilient transportation networks 面向弹性交通网络的动态路径冗余战略规划

IF 7.6 1区工程技术 Q1 TRANSPORTATION SCIENCE & TECHNOLOGY

Transportation Research Part C-Emerging Technologies

Pub Date : 2025-12-25 DOI: 10.1016/j.trc.2025.105503

Kai Qu , Xiangdong Xu , Anthony Chen

Sufficient route redundancy ensures the availability of alternative routes during disruptions for critical trips (e.g., evacuations, relief transportation), which is essential for sustaining transportation network resilience. Existing studies on route redundancy focus mainly on assessment rather than optimization, and most adopt a static perspective that ignores the inherently dynamic nature of resilience. This study introduces the definition, evaluation, and optimization methods for dynamic network redundancy. We propose period-based metrics that account for travelers’ adaptive behavior and network congestion and develop a link-based day-to-day model under uncertainty coupled with Dial counting to consistently measure redundancy. A two-stage stochastic bi-level programming model is then formulated to identify investment strategies that maximize expected dynamic redundancy under uncertain disruptions. In the first stage, planners allocate budgets between link retrofitting and new link construction, while in the second stage, travelers dynamically reroute on a daily basis following disruptions. To solve the resulting non-convex optimization problem, we implement a Bayesian optimization framework with parallel scenario evaluation. Experiments on both test and real-world networks demonstrate the properties, features, and applicability of the proposed methods. Results indicate that accounting for dynamic traffic evolution and congestion can reduce redundancy estimates by up to 40% compared to static assessments in the 16-node test network, particularly under severe disruptions and high congestion levels. The spatial–temporal evolution of congestion patterns, which influences travelers’ perception of alternative routes, is naturally captured by dynamic redundancy but overlooked in static assessments. In the Anaheim network, increasing the budget from $300 million to $1200 million raises dynamic redundancy from 4.5 % to 10.1 %, illustrating diminishing marginal returns. The framework developed in this study provides a decision-support tool for more informed, resilience-oriented network planning.

足够的路线冗余确保在关键行程（如疏散、救援运输）中断期间有替代路线的可用性，这对维持交通网络的弹性至关重要。现有的路径冗余研究主要侧重于评估而非优化，大多采用静态视角，忽略了弹性的内在动态性。介绍了动态网络冗余的定义、评价和优化方法。我们提出了基于周期的指标，考虑旅行者的自适应行为和网络拥塞，并在不确定性下开发了一个基于链路的日常模型，并结合拨号计数来一致地测量冗余。然后建立了一个两阶段的随机双级规划模型，以确定在不确定中断下最大化期望动态冗余的投资策略。在第一阶段，规划者在线路改造和新线路建设之间分配预算，而在第二阶段，旅客在交通中断后每天动态地改变路线。为了解决由此产生的非凸优化问题，我们实现了一个并行场景评估的贝叶斯优化框架。在测试和现实网络上的实验证明了所提出方法的性质、特征和适用性。结果表明，与16节点测试网络中的静态评估相比，考虑动态流量演变和拥塞可以将冗余估计减少多达40%，特别是在严重中断和高拥塞水平下。拥堵模式的时空演变会影响出行者对替代路线的感知，动态冗余很自然地捕捉到了这一点，但在静态评估中却被忽视了。在阿纳海姆网络中，将预算从3亿美元增加到12亿美元，将使动态冗余从4.5%增加到10.1%，说明边际收益递减。本研究开发的框架为更明智的、面向弹性的网络规划提供了决策支持工具。

{"title":"Dynamic route redundancy-oriented strategic planning towards resilient transportation networks","authors":"Kai Qu , Xiangdong Xu , Anthony Chen","doi":"10.1016/j.trc.2025.105503","DOIUrl":"10.1016/j.trc.2025.105503","url":null,"abstract":"<div><div>Sufficient route redundancy ensures the availability of alternative routes during disruptions for critical trips (e.g., evacuations, relief transportation), which is essential for sustaining transportation network resilience. Existing studies on route redundancy focus mainly on assessment rather than optimization, and most adopt a static perspective that ignores the inherently dynamic nature of resilience. This study introduces the definition, evaluation, and optimization methods for dynamic network redundancy. We propose period-based metrics that account for travelers’ adaptive behavior and network congestion and develop a link-based day-to-day model under uncertainty coupled with Dial counting to consistently measure redundancy. A two-stage stochastic bi-level programming model is then formulated to identify investment strategies that maximize expected dynamic redundancy under uncertain disruptions. In the first stage, planners allocate budgets between link retrofitting and new link construction, while in the second stage, travelers dynamically reroute on a daily basis following disruptions. To solve the resulting non-convex optimization problem, we implement a Bayesian optimization framework with parallel scenario evaluation. Experiments on both test and real-world networks demonstrate the properties, features, and applicability of the proposed methods. Results indicate that accounting for dynamic traffic evolution and congestion can reduce redundancy estimates by up to 40% compared to static assessments in the 16-node test network, particularly under severe disruptions and high congestion levels. The spatial–temporal evolution of congestion patterns, which influences travelers’ perception of alternative routes, is naturally captured by dynamic redundancy but overlooked in static assessments. In the Anaheim network, increasing the budget from $300 million to $1200 million raises dynamic redundancy from 4.5 % to 10.1 %, illustrating diminishing marginal returns. The framework developed in this study provides a decision-support tool for more informed, resilience-oriented network planning.</div></div>","PeriodicalId":54417,"journal":{"name":"Transportation Research Part C-Emerging Technologies","volume":"184 ","pages":"Article 105503"},"PeriodicalIF":7.6,"publicationDate":"2025-12-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145824128","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Traffic dynamics modeling with an extended S3 car-following model 使用扩展S3车辆跟随模型的交通动态建模

IF 7.6 1区工程技术 Q1 TRANSPORTATION SCIENCE & TECHNOLOGY

Transportation Research Part C-Emerging Technologies

Pub Date : 2025-12-24 DOI: 10.1016/j.trc.2025.105494

Zelin Wang , Yuqian Lin , Zhiyuan Liu , Yu Dong , Yuan Zheng , Pan Liu , Qixiu Cheng

In recent decades, there has been a notable surge in research attention on microscopic car-following modeling, reflecting its indispensable role in investigating and simulating human driving behavior. Breakthroughs in operations research and machine learning have reinvigorated scholarly interest in addressing intricate traffic flow issues encountered in real-world scenarios. This study advances the current state of traffic dynamics modeling by presenting a novel extended S-shaped three-parameter (ES3) car-following model, achieving a harmonious balance between empirical accuracy and mathematical simplicity while enabling macroscopic-microscopic integration. Both local and string stability conditions of the ES3 model are derived and corroborated by numerical experiments. Utilizing real-world and autonomous vehicle trajectory data, we calibrate and validate the ES3 model using three distinct methodologies: single-vehicle microscopic model calibration, multi-vehicle microscopic model calibration, and integrated self-consistent macroscopic-microscopic model calibration. These methodologies target different aspects, including individual vehicle behavior, platoon trajectories, and the integration of macroscopic and microscopic traffic flows. Comparative analyses against existing physics-based models demonstrate the exceptional performance of the proposed ES3 model in microscopic traffic flow modeling. Overall, the experimental results indicate its capability to accurately reproduce human drivers’ and autonomous vehicles’ car-following behaviors while elucidating the underlying mechanisms governing the observed macroscopic traffic phenomena.

近几十年来，微观汽车跟随建模的研究得到了极大的关注，反映了微观汽车跟随建模在研究和模拟人类驾驶行为中不可或缺的作用。运筹学和机器学习的突破重新激发了学术界对解决现实世界中遇到的复杂交通流问题的兴趣。本研究通过提出一种新的扩展s形三参数（ES3）车辆跟随模型，推进了交通动力学建模的现状，实现了经验准确性和数学简洁性之间的和谐平衡，同时实现了宏观与微观的整合。推导了ES3模型的局部稳定条件和弦稳定条件，并通过数值实验进行了验证。利用真实世界和自动驾驶车辆的轨迹数据，我们使用三种不同的方法校准和验证ES3模型：单车辆微观模型校准、多车辆微观模型校准和综合自一致宏观-微观模型校准。这些方法针对不同的方面，包括单个车辆行为，队列轨迹，以及宏观和微观交通流的整合。通过与现有物理模型的对比分析，证明了ES3模型在微观交通流建模中的卓越性能。总的来说，实验结果表明，它能够准确地再现人类驾驶员和自动驾驶汽车的跟车行为，同时阐明了控制观察到的宏观交通现象的潜在机制。

{"title":"Traffic dynamics modeling with an extended S3 car-following model","authors":"Zelin Wang , Yuqian Lin , Zhiyuan Liu , Yu Dong , Yuan Zheng , Pan Liu , Qixiu Cheng","doi":"10.1016/j.trc.2025.105494","DOIUrl":"10.1016/j.trc.2025.105494","url":null,"abstract":"<div><div>In recent decades, there has been a notable surge in research attention on microscopic car-following modeling, reflecting its indispensable role in investigating and simulating human driving behavior. Breakthroughs in operations research and machine learning have reinvigorated scholarly interest in addressing intricate traffic flow issues encountered in real-world scenarios. This study advances the current state of traffic dynamics modeling by presenting a novel extended S-shaped three-parameter (ES3) car-following model, achieving a harmonious balance between empirical accuracy and mathematical simplicity while enabling macroscopic-microscopic integration. Both local and string stability conditions of the ES3 model are derived and corroborated by numerical experiments. Utilizing real-world and autonomous vehicle trajectory data, we calibrate and validate the ES3 model using three distinct methodologies: single-vehicle microscopic model calibration, multi-vehicle microscopic model calibration, and integrated self-consistent macroscopic-microscopic model calibration. These methodologies target different aspects, including individual vehicle behavior, platoon trajectories, and the integration of macroscopic and microscopic traffic flows. Comparative analyses against existing physics-based models demonstrate the exceptional performance of the proposed ES3 model in microscopic traffic flow modeling. Overall, the experimental results indicate its capability to accurately reproduce human drivers’ and autonomous vehicles’ car-following behaviors while elucidating the underlying mechanisms governing the observed macroscopic traffic phenomena.</div></div>","PeriodicalId":54417,"journal":{"name":"Transportation Research Part C-Emerging Technologies","volume":"183 ","pages":"Article 105494"},"PeriodicalIF":7.6,"publicationDate":"2025-12-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145823187","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Transferability in data poisoning attacks on spatiotemporal traffic forecasting models 时空流量预测模型中数据投毒攻击的可移植性

IF 7.6 1区工程技术 Q1 TRANSPORTATION SCIENCE & TECHNOLOGY

Transportation Research Part C-Emerging Technologies

Pub Date : 2025-12-24 DOI: 10.1016/j.trc.2025.105501

Xin Wang , Feilong Wang , Yuan Hong , Xuegang (Jeff) Ban

Machine learning and deep learning models have been shown to be sensitive to adversarial perturbations within training data, where specially crafted examples can lead to compromised model performance. With the widespread adoption of deep learning models, such as graph neural networks (GNNs) and Transformers, for spatiotemporal traffic forecasting, the vulnerability of their training processes has emerged as a critical area for research. In this paper, we investigate the vulnerability of deep learning-based spatiotemporal traffic forecasting models during their training phase. We propose a practical, black-box poisoning attack framework to demonstrate the transferability of adversarial perturbations in data poisoning attacks on these forecasting models. Specifically, we craft adversarial perturbations against a public deep learning model on a public dataset and then apply these perturbations to other deep learning-based traffic forecasting models. Our transferred perturbations can universally compromise multiple traffic forecasting models without requiring access to any internal information or structure of the target models. Numerical experiments are conducted on six traffic state estimation and forecasting models across four traffic datasets. The revealed transferability highlights a substantial vulnerability in traffic forecasting models that are widely used in intelligent transportation applications.

机器学习和深度学习模型已被证明对训练数据中的对抗性扰动很敏感，其中精心制作的示例可能导致模型性能受损。随着深度学习模型（如图神经网络（gnn）和变形金刚）在时空交通预测中的广泛应用，其训练过程的脆弱性已成为一个重要的研究领域。本文研究了基于深度学习的时空交通预测模型在训练阶段的脆弱性。我们提出了一个实用的黑盒投毒攻击框架，以证明在这些预测模型的数据投毒攻击中对抗性扰动的可转移性。具体来说，我们针对公共数据集上的公共深度学习模型制作对抗性扰动，然后将这些扰动应用于其他基于深度学习的流量预测模型。我们的转移扰动可以普遍地破坏多个交通预测模型，而不需要访问目标模型的任何内部信息或结构。在4个交通数据集上对6种交通状态估计与预测模型进行了数值实验。揭示的可转移性突出了智能交通应用中广泛使用的交通预测模型的实质性漏洞。

{"title":"Transferability in data poisoning attacks on spatiotemporal traffic forecasting models","authors":"Xin Wang , Feilong Wang , Yuan Hong , Xuegang (Jeff) Ban","doi":"10.1016/j.trc.2025.105501","DOIUrl":"10.1016/j.trc.2025.105501","url":null,"abstract":"<div><div>Machine learning and deep learning models have been shown to be sensitive to adversarial perturbations within training data, where specially crafted examples can lead to compromised model performance. With the widespread adoption of deep learning models, such as graph neural networks (GNNs) and Transformers, for spatiotemporal traffic forecasting, the vulnerability of their training processes has emerged as a critical area for research. In this paper, we investigate the vulnerability of deep learning-based spatiotemporal traffic forecasting models during their training phase. We propose a practical, black-box poisoning attack framework to demonstrate the <em>transferability</em> of adversarial perturbations in data poisoning attacks on these forecasting models. Specifically, we craft adversarial perturbations against a public deep learning model on a public dataset and then apply these perturbations to other deep learning-based traffic forecasting models. Our transferred perturbations can universally compromise multiple traffic forecasting models without requiring access to any internal information or structure of the target models. Numerical experiments are conducted on six traffic state estimation and forecasting models across four traffic datasets. The revealed transferability highlights a substantial vulnerability in traffic forecasting models that are widely used in intelligent transportation applications.</div></div>","PeriodicalId":54417,"journal":{"name":"Transportation Research Part C-Emerging Technologies","volume":"183 ","pages":"Article 105501"},"PeriodicalIF":7.6,"publicationDate":"2025-12-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145823188","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Iterative physics-enhanced residual learning for context-aware traffic assignment under biased inputs 有偏差输入下环境感知交通分配的迭代物理增强残差学习

IF 7.6 1区工程技术 Q1 TRANSPORTATION SCIENCE & TECHNOLOGY

Transportation Research Part C-Emerging Technologies

Pub Date : 2025-12-23 DOI: 10.1016/j.trc.2025.105498

Yuxin Shi , Keke Long , William H.K. Lam , Xiaopeng Li , Wei Ma

Hybrid approaches combining physics-based models and data-driven methods have shown promise in traffic modeling by leveraging physical structure while enhancing learning flexibility. One representative example is Physics-Enhanced Residual Learning (PERL), which augments physics-based predictions with learned residuals to correct modeling errors. However, the effectiveness of physics-based model can degrade under biased input features. To address this limitation, we propose iterative Physics-Enhanced Residual Learning (iPERL), an end-to-end framework designed to improve the robustness of physics-guided models under biased inputs. We apply iPERL to context-aware traffic assignment, in which explanatory inputs such as OD demand, link and node characteristics (e.g., capacity, free-flow speed), and performance function parameters may be biased due to indirect observations and calibration errors, while traffic conditions simultaneously vary with contextual factors like time and weather. iPERL extends standard PERL by incorporating a residual-based input correction mechanism that iteratively calibrates these biased inputs using feedback from residuals between predicted and observed flows. By integrating contextual features, iPERL further enables adaptive correction strategies under diverse traffic scenarios. We evaluate the framework on both synthetic and real-world networks. Results show that iPERL consistently outperforms baseline methods, including standard PERL, particularly when input bias or data scarcity is present. The proposed framework offers a robust, interpretable, and data-efficient solution for traffic flow estimation, with potential for generalization across networks and practical applications.

结合基于物理的模型和数据驱动的方法的混合方法在利用物理结构的同时增强学习灵活性的交通建模中显示出了希望。一个典型的例子是物理增强残差学习（PERL），它通过学习残差来增强基于物理的预测，以纠正建模错误。然而，在有偏的输入特征下，基于物理的模型的有效性会下降。为了解决这一限制，我们提出了迭代物理增强残差学习（iPERL），这是一个端到端框架，旨在提高有偏差输入下物理引导模型的鲁棒性。我们将iPERL应用于上下文感知交通分配，其中解释性输入，如OD需求、链路和节点特征（如容量、自由流速度）和性能函数参数可能由于间接观测和校准误差而有偏差，而交通状况同时随着时间和天气等上下文因素而变化。iPERL通过结合基于残差的输入校正机制来扩展标准PERL，该机制使用来自预测流和观察流之间残差的反馈来迭代地校准这些有偏差的输入。通过集成上下文特性，iPERL进一步支持在不同流量场景下的自适应校正策略。我们在合成网络和现实世界网络上评估了该框架。结果表明，iPERL始终优于基准方法，包括标准PERL，特别是在存在输入偏差或数据稀缺的情况下。所提出的框架为交通流量估计提供了一个健壮的、可解释的、数据高效的解决方案，具有跨网络推广和实际应用的潜力。

{"title":"Iterative physics-enhanced residual learning for context-aware traffic assignment under biased inputs","authors":"Yuxin Shi , Keke Long , William H.K. Lam , Xiaopeng Li , Wei Ma","doi":"10.1016/j.trc.2025.105498","DOIUrl":"10.1016/j.trc.2025.105498","url":null,"abstract":"<div><div>Hybrid approaches combining physics-based models and data-driven methods have shown promise in traffic modeling by leveraging physical structure while enhancing learning flexibility. One representative example is Physics-Enhanced Residual Learning (PERL), which augments physics-based predictions with learned residuals to correct modeling errors. However, the effectiveness of physics-based model can degrade under biased input features. To address this limitation, we propose iterative Physics-Enhanced Residual Learning (iPERL), an end-to-end framework designed to improve the robustness of physics-guided models under biased inputs. We apply iPERL to context-aware traffic assignment, in which explanatory inputs such as OD demand, link and node characteristics (e.g., capacity, free-flow speed), and performance function parameters may be biased due to indirect observations and calibration errors, while traffic conditions simultaneously vary with contextual factors like time and weather. iPERL extends standard PERL by incorporating a residual-based input correction mechanism that iteratively calibrates these biased inputs using feedback from residuals between predicted and observed flows. By integrating contextual features, iPERL further enables adaptive correction strategies under diverse traffic scenarios. We evaluate the framework on both synthetic and real-world networks. Results show that iPERL consistently outperforms baseline methods, including standard PERL, particularly when input bias or data scarcity is present. The proposed framework offers a robust, interpretable, and data-efficient solution for traffic flow estimation, with potential for generalization across networks and practical applications.</div></div>","PeriodicalId":54417,"journal":{"name":"Transportation Research Part C-Emerging Technologies","volume":"183 ","pages":"Article 105498"},"PeriodicalIF":7.6,"publicationDate":"2025-12-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145823196","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0