Pub Date : 2026-01-02DOI: 10.1016/j.tre.2025.104636
Karolin Eisele, Alf Kimms
Natural disasters such as floods occur more and more frequently due to climate change and claim many victims. If protective measures such as floodplains and dams are not sufficient or are damaged, emergency services must be deployed. In order to be able to deploy them as effectively as possible, we present a model for emergency services planning in the event of flooding. The mathematical model is based on the idea that the area of interest is subdivided into cells and snapshots of the situation are considered at discrete time periods. This way, we can model the spread of water over time taking the specific profile of the terrain into account. Also, the locations and the movement of the emergency teams can be described with user–specified granularity. Since solving such models optimally is out of the scope of today’s computational capabilities, we discuss several variants of so–called construction heuristics. Such methods run fast and produce results that help to assess a flood situation and about what can be achieved over time by fighting the floods. Such insights may not only help after the occurrence of an event, but also in advance in order to be prepared better. In a computational study the performance of heuristics based in simple priority rules is studied.
{"title":"The flood fighting problem: A basic model and construction heuristics","authors":"Karolin Eisele, Alf Kimms","doi":"10.1016/j.tre.2025.104636","DOIUrl":"10.1016/j.tre.2025.104636","url":null,"abstract":"<div><div>Natural disasters such as floods occur more and more frequently due to climate change and claim many victims. If protective measures such as floodplains and dams are not sufficient or are damaged, emergency services must be deployed. In order to be able to deploy them as effectively as possible, we present a model for emergency services planning in the event of flooding. The mathematical model is based on the idea that the area of interest is subdivided into cells and snapshots of the situation are considered at discrete time periods. This way, we can model the spread of water over time taking the specific profile of the terrain into account. Also, the locations and the movement of the emergency teams can be described with user–specified granularity. Since solving such models optimally is out of the scope of today’s computational capabilities, we discuss several variants of so–called construction heuristics. Such methods run fast and produce results that help to assess a flood situation and about what can be achieved over time by fighting the floods. Such insights may not only help after the occurrence of an event, but also in advance in order to be prepared better. In a computational study the performance of heuristics based in simple priority rules is studied.</div></div>","PeriodicalId":49418,"journal":{"name":"Transportation Research Part E-Logistics and Transportation Review","volume":"208 ","pages":"Article 104636"},"PeriodicalIF":8.8,"publicationDate":"2026-01-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145886031","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}
Pub Date : 2025-12-31DOI: 10.1016/j.tre.2025.104643
Qi Wang , Yankui Liu , Guoqing Zhang
The rapid proliferation of electric vehicles (EVs) has led to a significant increase in the quantity of used electric vehicle batteries (EVBs). This necessitates the design of a waste reverse supply chain to reuse and recycle EVBs and protect the environment. This paper examines an integrated reuse network design and pricing problem for EVBs, which involves two stakeholders: an echelon utilization enterprise (leader) and a recycling company (follower). Two stakeholders interact through a hierarchical decision-making process under the uncertainty of return quantity. To tackle this problem, we present two bilevel globalized distributionally robust (GDR) design and pricing models. The leader optimizes the locations of collection and echelon utilization centers, the transportation of used EVBs, and pricing strategies to maximize profit. The follower determines the quantity of used EVBs to purchase for dismantling and recycling in order to maximize profit. We derive computationally tractable reformulations of GDR expectation and chance constraints using Lagrangian duality and conjugate function. To efficiently solve the resulting joint chance-constrained model, we propose a tailored branch-and-cut (B&C) algorithm incorporating a strengthened formulation. A real-world case study is conducted to validate the superiority of the proposed methods. Results demonstrate that the globalized distributionally robust optimization models exhibit greater robustness than stochastic optimization models. The computational performance of the tailored B&C algorithm incorporating a strengthened formulation is assessed compared to the standard solver. We also analyze the impact of globalized sensitivity parameter, Wasserstein radius, norm choice, and tolerance level on profitability and provide decision-makers with insights for choosing parameters.
{"title":"Robust design and pricing of electric vehicle battery reuse network by tailored branch-and-cut algorithm","authors":"Qi Wang , Yankui Liu , Guoqing Zhang","doi":"10.1016/j.tre.2025.104643","DOIUrl":"10.1016/j.tre.2025.104643","url":null,"abstract":"<div><div>The rapid proliferation of electric vehicles (EVs) has led to a significant increase in the quantity of used electric vehicle batteries (EVBs). This necessitates the design of a waste reverse supply chain to reuse and recycle EVBs and protect the environment. This paper examines an integrated reuse network design and pricing problem for EVBs, which involves two stakeholders: an echelon utilization enterprise (leader) and a recycling company (follower). Two stakeholders interact through a hierarchical decision-making process under the uncertainty of return quantity. To tackle this problem, we present two bilevel globalized distributionally robust (GDR) design and pricing models. The leader optimizes the locations of collection and echelon utilization centers, the transportation of used EVBs, and pricing strategies to maximize profit. The follower determines the quantity of used EVBs to purchase for dismantling and recycling in order to maximize profit. We derive computationally tractable reformulations of GDR expectation and chance constraints using Lagrangian duality and conjugate function. To efficiently solve the resulting joint chance-constrained model, we propose a tailored branch-and-cut (B&C) algorithm incorporating a strengthened formulation. A real-world case study is conducted to validate the superiority of the proposed methods. Results demonstrate that the globalized distributionally robust optimization models exhibit greater robustness than stochastic optimization models. The computational performance of the tailored B&C algorithm incorporating a strengthened formulation is assessed compared to the standard solver. We also analyze the impact of globalized sensitivity parameter, Wasserstein radius, norm choice, and tolerance level on profitability and provide decision-makers with insights for choosing parameters.</div></div>","PeriodicalId":49418,"journal":{"name":"Transportation Research Part E-Logistics and Transportation Review","volume":"207 ","pages":"Article 104643"},"PeriodicalIF":8.8,"publicationDate":"2025-12-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145885065","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}
Pub Date : 2025-12-30DOI: 10.1016/j.tre.2025.104641
Nan Zheng, Shukai Li, Yin Yuan, Dongfan Xie
The distribution of passenger demands on certain urban railway lines exhibits obvious spatiotemporal imbalances, posing challenges for the traditional fixed formation mode. This paper presents the optimization of the virtual formation train timetable and rolling stock utilization strategy, which aims to maximize the quantity of connections and minimize the number of detained passengers. A mixed-integer nonlinear programming model (MINLP) is formulated to characterize this problem, in which the coupling/decoupling operations between different types of rolling stock are considered. By applying linearization techniques, the aforementioned MINLP model can be transformed into a mixed-integer linear programming (MILP) model. To effectively address the model, a two-stage (TS) optimization approach is designed to decompose the original problem into two sequential steps for the solution. In the first stage, a reduced-scale optimization problem is solved, focusing solely on a subset of services; then, the partial binary variables obtained from the first stage are incorporated into the original problem for further resolution in the second stage. Furthermore, we design an accelerated technique of bound contraction based on logical inference to enhance the solving efficiency of the second stage. Five sets of numerical experiments based on the Beijing metro Yizhuang line are conducted to verify the effectiveness and practicability of the model and algorithm. The experimental results illustrate that the virtual formation mode can effectively address the spatiotemporal imbalances of passenger demands on the line. The proposed TS approach is also proven to exhibit greater efficiency than traditional heuristic algorithms, such as genetic algorithm (GA), for large-scale problems.
{"title":"Train timetable optimization for urban railway systems under the virtual formation mode combined with the rolling stock utilization strategy","authors":"Nan Zheng, Shukai Li, Yin Yuan, Dongfan Xie","doi":"10.1016/j.tre.2025.104641","DOIUrl":"10.1016/j.tre.2025.104641","url":null,"abstract":"<div><div>The distribution of passenger demands on certain urban railway lines exhibits obvious spatiotemporal imbalances, posing challenges for the traditional fixed formation mode. This paper presents the optimization of the virtual formation train timetable and rolling stock utilization strategy, which aims to maximize the quantity of connections and minimize the number of detained passengers. A mixed-integer nonlinear programming model (MINLP) is formulated to characterize this problem, in which the coupling/decoupling operations between different types of rolling stock are considered. By applying linearization techniques, the aforementioned MINLP model can be transformed into a mixed-integer linear programming (MILP) model. To effectively address the model, a two-stage (TS) optimization approach is designed to decompose the original problem into two sequential steps for the solution. In the first stage, a reduced-scale optimization problem is solved, focusing solely on a subset of services; then, the partial binary variables obtained from the first stage are incorporated into the original problem for further resolution in the second stage. Furthermore, we design an accelerated technique of bound contraction based on logical inference to enhance the solving efficiency of the second stage. Five sets of numerical experiments based on the Beijing metro Yizhuang line are conducted to verify the effectiveness and practicability of the model and algorithm. The experimental results illustrate that the virtual formation mode can effectively address the spatiotemporal imbalances of passenger demands on the line. The proposed TS approach is also proven to exhibit greater efficiency than traditional heuristic algorithms, such as genetic algorithm (GA), for large-scale problems.</div></div>","PeriodicalId":49418,"journal":{"name":"Transportation Research Part E-Logistics and Transportation Review","volume":"207 ","pages":"Article 104641"},"PeriodicalIF":8.8,"publicationDate":"2025-12-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145885018","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}
Pub Date : 2025-12-30DOI: 10.1016/j.tre.2025.104594
Wei Xu , Zhixiao Wang , Zhenjie Zheng , Zhengli Wang , Hai Yang
The integration of drones with trucks or public transportation (PT) vehicles has become an increasingly popular strategy to extend the operational range of drone-based deliveries. Compared to truck-drone systems, PT-drone integration leverages existing public vehicles (e.g., buses) without the need for additional ground fleets, thereby reducing operational costs and environmental impact. However, existing studies on PT-drone integration have primarily focused on one-way parcel delivery tasks, whereas the simultaneous pickup and delivery (SPD) service remains underexplored. In this study, we develop a mixed integer linear programming (MILP) model that enables the effective synchronization of drone-based SPD service with fixed PT timetables and routes. Specifically, we first construct a time-expanded graph that encodes the spatial distribution of PT stations and the temporal scheduling of their associated trips across different lines. To capture the operational dynamics of drone-based SPD, we then formulate energy consumption as a function of flight time and payload, both of which evolve with routing decisions. Finally, the MILP model is solved to minimize both service time and system cost while ensuring compliance with operational constraints. We derive a set of valid inequalities to tighten the MILP formulation and enhance its overall computational efficiency. For large-scale instances, we also design a tailored Adaptive Large Neighborhood Search (ALNS) algorithm with problem-specific operators. Numerical experiments using real-world data from Nanjing, China, demonstrate the effectiveness of our proposed model in realizing the long-range SPD. The valid inequalities reduce the MILP solver time by 69.15 %, and the ALNS algorithm produces near-optimal solutions within reasonable time.
{"title":"Integrated routing of drones and public transportation vehicles for simultaneous parcel pickup and delivery","authors":"Wei Xu , Zhixiao Wang , Zhenjie Zheng , Zhengli Wang , Hai Yang","doi":"10.1016/j.tre.2025.104594","DOIUrl":"10.1016/j.tre.2025.104594","url":null,"abstract":"<div><div>The integration of drones with trucks or public transportation (PT) vehicles has become an increasingly popular strategy to extend the operational range of drone-based deliveries. Compared to truck-drone systems, PT-drone integration leverages existing public vehicles (e.g., buses) without the need for additional ground fleets, thereby reducing operational costs and environmental impact. However, existing studies on PT-drone integration have primarily focused on one-way parcel delivery tasks, whereas the simultaneous pickup and delivery (SPD) service remains underexplored. In this study, we develop a mixed integer linear programming (MILP) model that enables the effective synchronization of drone-based SPD service with fixed PT timetables and routes. Specifically, we first construct a time-expanded graph that encodes the spatial distribution of PT stations and the temporal scheduling of their associated trips across different lines. To capture the operational dynamics of drone-based SPD, we then formulate energy consumption as a function of flight time and payload, both of which evolve with routing decisions. Finally, the MILP model is solved to minimize both service time and system cost while ensuring compliance with operational constraints. We derive a set of valid inequalities to tighten the MILP formulation and enhance its overall computational efficiency. For large-scale instances, we also design a tailored Adaptive Large Neighborhood Search (ALNS) algorithm with problem-specific operators. Numerical experiments using real-world data from Nanjing, China, demonstrate the effectiveness of our proposed model in realizing the long-range SPD. The valid inequalities reduce the MILP solver time by 69.15 %, and the ALNS algorithm produces near-optimal solutions within reasonable time.</div></div>","PeriodicalId":49418,"journal":{"name":"Transportation Research Part E-Logistics and Transportation Review","volume":"207 ","pages":"Article 104594"},"PeriodicalIF":8.8,"publicationDate":"2025-12-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145885066","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}
Pub Date : 2025-12-30DOI: 10.1016/j.tre.2025.104588
Jeongin Yun , Seungmin Oh , Jinwoo Lee
Connected and autonomous vehicle (CAV) platooning, where a group of CAVs travel closely together at higher speeds, has the potential to improve both traffic capacity and free-flow speed of mixed traffic on roads. In this paper, we present a dedicated lane management framework based on an analytical understanding of mixed traffic involving CAVs and human-driven vehicles (HDVs), taking into account diverse headways, free-flow speeds, and CAV penetration rates. This framework is a bi-criteria optimization that maximizes both traffic capacity and free-flow time-mean speed of a multi-lane section, where each lane can be a non-dedicated lane, a CAV-dedicated lane, or an HDV-dedicated lane. In the capacity-maximizing case, through using both types of dedicated lanes, our approach can consistently maximize capacity across various environmental settings, such as lane numbers, CAV rates, and car-following aggressiveness. The optimal dedicated lane management scheme is summarized as follows: implement HDV-dedicated lane(s) when the total CAV ratio is low, and introduce CAV-dedicated lane(s) otherwise. The scheme aims to consolidate CAVs as much as possible to maximize the number of platooning events. In the capacity-and-speed-maximizing case, CAV-dedicated lane(s) are introduced at lower CAV penetration rates compared to the capacity-maximizing case, with greater emphasis on speed, resulting in more complete separation between CAVs and HDVs. In the bi-criteria optimization, a Pareto solution set is found, illustrating the tradeoff between two objectives, which allows transportation planners flexibility in selecting lane management strategies in accordance with operational priorities. Finally, we validate the proposed framework through agent-based simulations in VISSIM, demonstrating its effectiveness.
{"title":"Optimal dedicated lane management for mixed traffic with connected and autonomous vehicles accounting for heterogeneous headways and speeds","authors":"Jeongin Yun , Seungmin Oh , Jinwoo Lee","doi":"10.1016/j.tre.2025.104588","DOIUrl":"10.1016/j.tre.2025.104588","url":null,"abstract":"<div><div>Connected and autonomous vehicle (CAV) platooning, where a group of CAVs travel closely together at higher speeds, has the potential to improve both traffic capacity and free-flow speed of mixed traffic on roads. In this paper, we present a dedicated lane management framework based on an analytical understanding of mixed traffic involving CAVs and human-driven vehicles (HDVs), taking into account diverse headways, free-flow speeds, and CAV penetration rates. This framework is a bi-criteria optimization that maximizes both traffic capacity and free-flow time-mean speed of a multi-lane section, where each lane can be a non-dedicated lane, a CAV-dedicated lane, or an HDV-dedicated lane. In the capacity-maximizing case, through using both types of dedicated lanes, our approach can consistently maximize capacity across various environmental settings, such as lane numbers, CAV rates, and car-following aggressiveness. The optimal dedicated lane management scheme is summarized as follows: implement HDV-dedicated lane(s) when the total CAV ratio is low, and introduce CAV-dedicated lane(s) otherwise. The scheme aims to consolidate CAVs as much as possible to maximize the number of platooning events. In the capacity-and-speed-maximizing case, CAV-dedicated lane(s) are introduced at lower CAV penetration rates compared to the capacity-maximizing case, with greater emphasis on speed, resulting in more complete separation between CAVs and HDVs. In the bi-criteria optimization, a Pareto solution set is found, illustrating the tradeoff between two objectives, which allows transportation planners flexibility in selecting lane management strategies in accordance with operational priorities. Finally, we validate the proposed framework through agent-based simulations in VISSIM, demonstrating its effectiveness.</div></div>","PeriodicalId":49418,"journal":{"name":"Transportation Research Part E-Logistics and Transportation Review","volume":"207 ","pages":"Article 104588"},"PeriodicalIF":8.8,"publicationDate":"2025-12-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145885017","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}
Pub Date : 2025-12-30DOI: 10.1016/j.tre.2025.104615
Mohammad Mahdi Vali-Siar , Hamid Tikani , Emrah Demir , Yousof Shamstabar
During large-scale disruptions, particularly super-disruptions such as global pandemics or large-scale natural disasters, supply chains are exposed to significant adverse impacts. This paper addresses the resilience in a supply chain network design problem under disruption risk by explicitly modeling the dependency between the inter-arrival times of disruptive events and severity of their consequences. A novel data-driven stochastic optimization framework is proposed to consider the ripple effects that typically propagate across supply chain networks following severe disruptions. Specifically, we have devised a hybrid methodology that integrates a clustering algorithm (unsupervised machine learning technique), a phase-type disruption model, and a two-stage stochastic model. To elaborate, a genetic-based clustering algorithm is used to identify the structure dependencies in the input data. Phase-type distributions and their associated theorems are then used to determine the probability distributions of disruptions. A novel mathematical model is developed to design the supply chain using the scenarios generated based on the obtained distributions, which is then solved using the Lagrangian decomposition combined with a new hyper-matheuristic algorithm. The computational efficiency and practical value of the proposed approach are demonstrated through a real-world case study. The findings highlight the effectiveness of developed methodology in designing a resilient supply chain, the proposed resilience strategies substantially improve the supply chain’s performance compared to a non-resilient approach.
{"title":"Resilient supply chain network design under super-disruption considering inter-arrival time dependency: a new data-driven stochastic optimization approach","authors":"Mohammad Mahdi Vali-Siar , Hamid Tikani , Emrah Demir , Yousof Shamstabar","doi":"10.1016/j.tre.2025.104615","DOIUrl":"10.1016/j.tre.2025.104615","url":null,"abstract":"<div><div>During large-scale disruptions, particularly super-disruptions such as global pandemics or large-scale natural disasters, supply chains are exposed to significant adverse impacts. This paper addresses the resilience in a supply chain network design problem under disruption risk by explicitly modeling the dependency between the inter-arrival times of disruptive events and severity of their consequences. A novel data-driven stochastic optimization framework is proposed to consider the ripple effects that typically propagate across supply chain networks following severe disruptions. Specifically, we have devised a hybrid methodology that integrates a clustering algorithm (unsupervised machine learning technique), a phase-type disruption model, and a two-stage stochastic model. To elaborate, a genetic-based clustering algorithm is used to identify the structure dependencies in the input data. Phase-type distributions and their associated theorems are then used to determine the probability distributions of disruptions. A novel mathematical model is developed to design the supply chain using the scenarios generated based on the obtained distributions, which is then solved using the Lagrangian decomposition combined with a new hyper-matheuristic algorithm. The computational efficiency and practical value of the proposed approach are demonstrated through a real-world case study. The findings highlight the effectiveness of developed methodology in designing a resilient supply chain, the proposed resilience strategies substantially improve the supply chain’s performance compared to a non-resilient approach.</div></div>","PeriodicalId":49418,"journal":{"name":"Transportation Research Part E-Logistics and Transportation Review","volume":"207 ","pages":"Article 104615"},"PeriodicalIF":8.8,"publicationDate":"2025-12-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145885064","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}
Pub Date : 2025-12-28DOI: 10.1016/j.tre.2025.104607
Irfan Ullah , Muhammad Asim Ayaz , Minghui Zhong , Xiyan Mao , Quan Yuan
Logistics Electric Vehicles (LEVs) are increasingly essential for sustainable urban freight and last-mile delivery, driven by the global push toward low-emission transportation and smart mobility ecosystems. However, optimizing LEV operations remains challenging due to the complex interplay of energy constraints, charging behavior, and urban logistics dynamics. This study aims to predict the dwell time (i.e., stop duration) of LEVs using an interpretable machine learning (ML) technique to support efficient fleet scheduling and energy planning. This study utilizes a real-world dataset of 1,065 LEV stops collected over one month in Shanghai, comprising operational, temporal, and energy-related variables. A stacked ensemble model integrating XGBoost, LightGBM, and CatBoost is developed to achieve high predictive accuracy, while SHAP analysis is employed to interpret the influence of key features. The proposed model achieves an R2 of 0.993, significantly outperforming individual learners, and reveals complex non-linear relationships among operational, temporal, and energy-related variables. SHAP analysis reveals that end state-of-charge (end_soc) and start_soc emerge as dominant drivers of dwell time, followed by trip speed, distance, time_of_day, and charging status indicators. These findings highlight the critical role of energy conditions and time windows in shaping dwell time. The study provides actionable insights for logistics firms, such as improved route optimization, charging station placement, and shift planning. It also offers policy guidance for urban planners and regulators in designing smart grid-compatible infrastructure, incentive schemes, and public–private data collaborations to enhance LEV ecosystem performance.
{"title":"Predicting dwell time of logistics electric vehicles in urban last-mile delivery: A SHAP-based ensemble approach","authors":"Irfan Ullah , Muhammad Asim Ayaz , Minghui Zhong , Xiyan Mao , Quan Yuan","doi":"10.1016/j.tre.2025.104607","DOIUrl":"10.1016/j.tre.2025.104607","url":null,"abstract":"<div><div>Logistics Electric Vehicles (LEVs) are increasingly essential for sustainable urban freight and last-mile delivery, driven by the global push toward low-emission transportation and smart mobility ecosystems. However, optimizing LEV operations remains challenging due to the complex interplay of energy constraints, charging behavior, and urban logistics dynamics. This study aims to predict the dwell time (i.e., stop duration) of LEVs using an interpretable machine learning (ML) technique to support efficient fleet scheduling and energy planning. This study utilizes a real-world dataset of 1,065 LEV stops collected over one month in Shanghai, comprising operational, temporal, and energy-related variables. A stacked ensemble model integrating XGBoost, LightGBM, and CatBoost is developed to achieve high predictive accuracy, while SHAP analysis is employed to interpret the influence of key features. The proposed model achieves an R<sup>2</sup> of 0.993, significantly outperforming individual learners, and reveals complex non-linear relationships among operational, temporal, and energy-related variables. SHAP analysis reveals that end state-of-charge (end_soc) and start_soc emerge as dominant drivers of dwell time, followed by trip speed, distance, time_of_day, and charging status indicators. These findings highlight the critical role of energy conditions and time windows in shaping dwell time. The study provides actionable insights for logistics firms, such as improved route optimization, charging station placement, and shift planning. It also offers policy guidance for urban planners and regulators in designing smart grid-compatible infrastructure, incentive schemes, and public–private data collaborations to enhance LEV ecosystem performance.</div></div>","PeriodicalId":49418,"journal":{"name":"Transportation Research Part E-Logistics and Transportation Review","volume":"207 ","pages":"Article 104607"},"PeriodicalIF":8.8,"publicationDate":"2025-12-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145885067","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}
Pub Date : 2025-12-27DOI: 10.1016/j.tre.2025.104647
Liang Zhao , Zhenggang He
International maritime transportation is a major yet complex source of greenhouse-gas emissions, whose systemic drivers and network formation mechanisms are not fully captured by existing, often isolated, methodologies. To bridge this gap, this study develops a multi-scale, integrated analytical framework. We first employ an environmentally extended multi-region input–output model to quantify global maritime embedded carbon flows (2000–2020). We then combine a high-precision machine-learning model (MLP) with SHapley Additive exPlanations (SHAP) analysis to identify key drivers, and finally apply a weighted exponential random-graph model to uncover network generative mechanisms. Our analysis yields three pivotal insights that offer new perspectives beyond conventional approaches: (1) The global flow network exhibits a polarized core–periphery structure centered on major hubs like China, Singapore, and the United States. (2) Bilateral flow intensity is primarily driven by asymmetric economic structures, operating through robust nonlinear (e.g., U-shaped, inverted U-shaped) channels rather than linear relationships. (3) Network formation is co-driven by homophily in consumption and heterophily in industrial structure, with geographic distance a persistent barrier. These findings directly inform international climate policy: they advocate for expanding emission responsibility to include major consumer nations and logistics hubs, and call for policies that account for the nonlinear, structural drivers of carbon exchange. The machine learning code and data have been uploaded to GitHub. URL: https://github.com/zhaoliangovo/Project-of-global-maritime-embedded-carbon-flow-network.
{"title":"Global maritime embedded carbon flow network: Key factors and formation mechanism","authors":"Liang Zhao , Zhenggang He","doi":"10.1016/j.tre.2025.104647","DOIUrl":"10.1016/j.tre.2025.104647","url":null,"abstract":"<div><div>International maritime transportation is a major yet complex source of greenhouse-gas emissions, whose systemic drivers and network formation mechanisms are not fully captured by existing, often isolated, methodologies. To bridge this gap, this study develops a multi-scale, integrated analytical framework. We first employ an environmentally extended multi-region input–output model to quantify global maritime embedded carbon flows (2000–2020). We then combine a high-precision machine-learning model (MLP) with SHapley Additive exPlanations (SHAP) analysis to identify key drivers, and finally apply a weighted exponential random-graph model to uncover network generative mechanisms. Our analysis yields three pivotal insights that offer new perspectives beyond conventional approaches: (1) The global flow network exhibits a polarized core–periphery structure centered on major hubs like China, Singapore, and the United States. (2) Bilateral flow intensity is primarily driven by asymmetric economic structures, operating through robust nonlinear (e.g., U-shaped, inverted U-shaped) channels rather than linear relationships. (3) Network formation is co-driven by homophily in consumption and heterophily in industrial structure, with geographic distance a persistent barrier. These findings directly inform international climate policy: they advocate for expanding emission responsibility to include major consumer nations and logistics hubs, and call for policies that account for the nonlinear, structural drivers of carbon exchange. The machine learning code and data have been uploaded to GitHub. URL: <span><span>https://github.com/zhaoliangovo/Project-of-global-maritime-embedded-carbon-flow-network</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":49418,"journal":{"name":"Transportation Research Part E-Logistics and Transportation Review","volume":"207 ","pages":"Article 104647"},"PeriodicalIF":8.8,"publicationDate":"2025-12-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145841515","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}
Pub Date : 2025-12-27DOI: 10.1016/j.tre.2025.104633
Menghang Wang, Lan Lu, Lindong Liu, Jie Wu
This paper considers a two-stage resource allocation game within a cooperative game framework from a platform perspective, where the customers’ demands are uncertain. To incentivize all customers (players) into the grand coalition for joint cost sharing in resource allocation, a critical issue for the platform is determining a fair and robust cost allocation solution. To address the challenge, we introduce the concept of the strict robust core to the operations research (OR) game with constraints and propose the Two-stage Resource Allocation-Robust Cost Sharing Problem (TRA-RCSP). Our approach integrates distributionally robust optimization (DRO) and distributionally favorable optimization (DFO) to improve computational tractability. By leveraging the polyhedral ambiguity set to model demand uncertainty, we calculate the worst-case cost for grand coalition and the best-case costs for subcoalitions. Additionally, we develop an iterative constraint generation algorithm to mitigate the exponential growth of constraints in TRA-RCSP. Numerical experiments demonstrate that our algorithm achieves excellent computational efficiency and the strict robust core significantly outperforms the cost allocation of SAA model across both robustness performance metrics, ensuring the formation of the grand cooperation and its long-term stability under uncertain demands.
{"title":"Cost allocation in a robust two-stage resource allocation game: Fairness and robustness","authors":"Menghang Wang, Lan Lu, Lindong Liu, Jie Wu","doi":"10.1016/j.tre.2025.104633","DOIUrl":"10.1016/j.tre.2025.104633","url":null,"abstract":"<div><div>This paper considers a two-stage resource allocation game within a cooperative game framework from a platform perspective, where the customers’ demands are uncertain. To incentivize all customers (players) into the grand coalition for joint cost sharing in resource allocation, a critical issue for the platform is determining a fair and robust cost allocation solution. To address the challenge, we introduce the concept of the strict robust core to the operations research (OR) game with constraints and propose the Two-stage Resource Allocation-Robust Cost Sharing Problem (TRA-RCSP). Our approach integrates distributionally robust optimization (DRO) and distributionally favorable optimization (DFO) to improve computational tractability. By leveraging the polyhedral ambiguity set to model demand uncertainty, we calculate the worst-case cost for grand coalition and the best-case costs for subcoalitions. Additionally, we develop an iterative constraint generation algorithm to mitigate the exponential growth of constraints in TRA-RCSP. Numerical experiments demonstrate that our algorithm achieves excellent computational efficiency and the strict robust core significantly outperforms the cost allocation of SAA model across both robustness performance metrics, ensuring the formation of the grand cooperation and its long-term stability under uncertain demands.</div></div>","PeriodicalId":49418,"journal":{"name":"Transportation Research Part E-Logistics and Transportation Review","volume":"207 ","pages":"Article 104633"},"PeriodicalIF":8.8,"publicationDate":"2025-12-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145845136","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}
Pub Date : 2025-12-27DOI: 10.1016/j.tre.2025.104648
Shangsong Long , Hui Zhao , Dan Zhu
Recent years have witnessed the growing influence of emerging digital port policies (DPPs) on the maritime shipping industry. These policies have had mixed impacts on firms’ profitability, with the effects varying according to their ESG (Environmental, Social, and Governance) profiles. Motivated by this, this study investigates how DPPs affect port and carrier profits under different ESG conditions. We employ two main approaches: (1) a government–firm evolutionary game theoretical framework with ESG heterogeneity, and (2) empirical analysis based on event study methodology. The theoretical models reveal optimal DPP strategies for governments and identify two stable ESG-related equilibria under varying market scenarios. Key factors such as tax rate and cost coefficient are found to influence equilibrium outcomes. Empirically, we analyze 129 listed maritime firms in China’s A-share market and compare two representative DPPs issued in 2019 and 2023. Results show that the 2019 policy had a negative impact on firm profitability, whereas the 2023 policy produced a positive effect. This phenomenon is consistent with the predictions derived from our theoretical models. Finally, we find that a firm’s ESG level plays a positive moderating role in the relationship between DPPs and firm performance. These findings can provide useful implications for the development and refinement of digital port policies.
{"title":"Transforming maritime supply chains through digital port policies considering ESG: an evolutionary game theoretical framework with empirical analysis","authors":"Shangsong Long , Hui Zhao , Dan Zhu","doi":"10.1016/j.tre.2025.104648","DOIUrl":"10.1016/j.tre.2025.104648","url":null,"abstract":"<div><div>Recent years have witnessed the growing influence of emerging digital port policies (DPPs) on the maritime shipping industry. These policies have had mixed impacts on firms’ profitability, with the effects varying according to their ESG (Environmental, Social, and Governance) profiles. Motivated by this, this study investigates how DPPs affect port and carrier profits under different ESG conditions. We employ two main approaches: (1) a government–firm evolutionary game theoretical framework with ESG heterogeneity, and (2) empirical analysis based on event study methodology. The theoretical models reveal optimal DPP strategies for governments and identify two stable ESG-related equilibria under varying market scenarios. Key factors such as tax rate and cost coefficient are found to influence equilibrium outcomes. Empirically, we analyze 129 listed maritime firms in China’s A-share market and compare two representative DPPs issued in 2019 and 2023. Results show that the 2019 policy had a negative impact on firm profitability, whereas the 2023 policy produced a positive effect. This phenomenon is consistent with the predictions derived from our theoretical models. Finally, we find that a firm’s ESG level plays a positive moderating role in the relationship between DPPs and firm performance. These findings can provide useful implications for the development and refinement of digital port policies.</div></div>","PeriodicalId":49418,"journal":{"name":"Transportation Research Part E-Logistics and Transportation Review","volume":"207 ","pages":"Article 104648"},"PeriodicalIF":8.8,"publicationDate":"2025-12-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145842054","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}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号