Pub Date : 2024-12-13DOI: 10.1016/j.tre.2024.103931
Mohammad Amin Amani, Samuel Asumadu Sarkodie, Jiuh-Biing Sheu, Mohammad Mahdi Nasiri, Reza Tavakkoli-Moghaddam
The incorporation of artificial intelligence (AI) and robust optimization methods for the planning and design of relief logistics networks under relief demand–supply uncertainty appears promising for intelligent disaster management (IDM). This research proposes a data-driven hybrid scenario-based robust (SBR) method for a mixed integer second-order cone programming (MISOCP) model that integrates machine learning with a hybrid robust optimization approach to address the above issue. A machine learning technique is utilized to cluster the casualties based on location coordinates and injury severity score. Moreover, the hybrid SBR optimization method and robust optimization based on the uncertainty sets technique are utilized to cope with uncertain parameters such as the probability of facility disruption, the number of wounded individuals, transportation time, and relief demand. Additionally, the epsilon-constraint technique is applied to seek the solution for the bi-objective model. Focusing on a real case (the Kermanshah disaster), our analytical results have demonstrated not only the validity but also the relative merits of the proposed methodology against typical stochastic and robust optimization approaches. Besides, the proposed method shows all casualties can be efficiently transported to receive medical services at a fair cost, which is crucial for disaster management.
{"title":"A data-driven hybrid scenario-based robust optimization method for relief logistics network design","authors":"Mohammad Amin Amani, Samuel Asumadu Sarkodie, Jiuh-Biing Sheu, Mohammad Mahdi Nasiri, Reza Tavakkoli-Moghaddam","doi":"10.1016/j.tre.2024.103931","DOIUrl":"https://doi.org/10.1016/j.tre.2024.103931","url":null,"abstract":"The incorporation of artificial intelligence (AI) and robust optimization methods for the planning and design of relief logistics networks under relief demand–supply uncertainty appears promising for intelligent disaster management (IDM). This research proposes a data-driven hybrid scenario-based robust (SBR) method for a mixed integer second-order cone programming (MISOCP) model that integrates machine learning with a hybrid robust optimization approach to address the above issue. A machine learning technique is utilized to cluster the casualties based on location coordinates and injury severity score. Moreover, the hybrid SBR optimization method and robust optimization based on the uncertainty sets technique are utilized to cope with uncertain parameters such as the probability of facility disruption, the number of wounded individuals, transportation time, and relief demand. Additionally, the epsilon-constraint technique is applied to seek the solution for the bi-objective model. Focusing on a real case (the Kermanshah disaster), our analytical results have demonstrated not only the validity but also the relative merits of the proposed methodology against typical stochastic and robust optimization approaches. Besides, the proposed method shows all casualties can be efficiently transported to receive medical services at a fair cost, which is crucial for disaster management.","PeriodicalId":49418,"journal":{"name":"Transportation Research Part E-Logistics and Transportation Review","volume":"5 1","pages":""},"PeriodicalIF":10.6,"publicationDate":"2024-12-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142816525","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 : 2024-12-12DOI: 10.1016/j.tre.2024.103901
Zhuolin Wang, Rongping Zhu, Jian-Ya Ding, Yu Yang, Keyou You
This work is concerned with the daily package shipment problem that aims to find low-cost paths for a large volume of packages and transportation vehicles over a network of transshipment centers (TCs). For Chinese courier companies, this typically involves tens of thousands of origin–destination (OD) pairs and has to be solved in a limited time window every early morning. Inspired by their industry practices, where most vehicles (99.8% for our industry partner STO) only unload packages after departing from the origin and the shipment volumes can be split, we propose a novel Localized Package Shipment with Partial Outsourcing (LPSPO) model for each TC to individually decide their daily shipment profiles, which aligns with their current operations. Though the number of OD pairs in our localized model is considerably reduced, it is strongly NP-hard and we exploit the model structure to design a column generation-based algorithm, which iteratively identifies profitable paths for the restricted master problem. Then, we develop problem-specific cutting planes and variable bound tightening techniques to accelerate our algorithm. An extensive numerical study validates that our algorithm significantly outperforms CPLEX in solving the LPSPO model. Finally, experiments on realistic instances from a leading Chinese courier company illustrate that the LPSPO model may reduce its transportation costs by up to 10 million USD annually.
{"title":"Localized package shipment with partial outsourcing: An exact optimization approach for Chinese courier companies","authors":"Zhuolin Wang, Rongping Zhu, Jian-Ya Ding, Yu Yang, Keyou You","doi":"10.1016/j.tre.2024.103901","DOIUrl":"https://doi.org/10.1016/j.tre.2024.103901","url":null,"abstract":"This work is concerned with the daily package shipment problem that aims to find low-cost paths for a large volume of packages and transportation vehicles over a network of transshipment centers (TCs). For Chinese courier companies, this typically involves tens of thousands of origin–destination (OD) pairs and has to be solved in a limited time window every early morning. Inspired by their industry practices, where most vehicles (99.8% for our industry partner STO) only unload packages after departing from the origin and the shipment volumes can be split, we propose a novel Localized Package Shipment with Partial Outsourcing (LPSPO) model for each TC to individually decide their daily shipment profiles, which aligns with their current operations. Though the number of OD pairs in our localized model is considerably reduced, it is <ce:italic>strongly NP-hard</ce:italic> and we exploit the model structure to design a column generation-based algorithm, which iteratively identifies profitable paths for the restricted master problem. Then, we develop problem-specific cutting planes and variable bound tightening techniques to accelerate our algorithm. An extensive numerical study validates that our algorithm significantly outperforms CPLEX in solving the LPSPO model. Finally, experiments on realistic instances from a leading Chinese courier company illustrate that the LPSPO model may reduce its transportation costs by up to 10 million USD annually.","PeriodicalId":49418,"journal":{"name":"Transportation Research Part E-Logistics and Transportation Review","volume":"118 1","pages":""},"PeriodicalIF":10.6,"publicationDate":"2024-12-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142816528","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 : 2024-12-12DOI: 10.1016/j.tre.2024.103903
Jie Wang, Benedict Jun Ma, Yanyi Yang, Chun-Hung Cheng, Yong-Hong Kuo
In group buying (GB), the retailer launches a deal with a discounted product price and a minimum group size requirement. Strategic consumers then determine whether to sign up for the GB deal or purchase the product at the regular price immediately. If GB fails, disappointed GB participants perceive a negative psychological utility and decide whether or not to buy it again at the regular price. Considering the disappointment caused by a GB failure, in our basic model, we formulate a two-period game to study the retailer’s optimal pricing decisions and consumers’ purchasing strategies. By deriving the likelihood function of a consumer signing up for the GB deal and utilizing rational expectations theory, we characterize how consumers form their own beliefs about the GB success rate. We find that consumer sentiment toward a failed GB deal serves an important role in a GB deal. Specifically, there exists a threshold of the disappointment factor where the retailer’s profit and consumers’ purchase decisions may change, and consumer leakage and retention may occur. We prove the existence of the subgame perfect Nash equilibrium and outline how the retailer should design the GB schedule. Our study shows that if properly designed, GB is an effective strategy to enhance consumer interests and improve profit. Moreover, a big discount should be offered when the disappointment factor is significant. When the consumer sentiment toward a failed GB deal is insignificant, the retailer should launch a GB deal; otherwise, he should provide the regular sales channel only. We conduct numerical experiments to study the impacts of different factors in a GB deal. Our key results continue to hold in several extensions from our basic model: retailer competition, observable secured group size, and social interactions between consumers.
{"title":"Group buying with consumer disappointment at failed deals","authors":"Jie Wang, Benedict Jun Ma, Yanyi Yang, Chun-Hung Cheng, Yong-Hong Kuo","doi":"10.1016/j.tre.2024.103903","DOIUrl":"https://doi.org/10.1016/j.tre.2024.103903","url":null,"abstract":"In group buying (GB), the retailer launches a deal with a discounted product price and a minimum group size requirement. Strategic consumers then determine whether to sign up for the GB deal or purchase the product at the regular price immediately. If GB fails, disappointed GB participants perceive a negative psychological utility and decide whether or not to buy it again at the regular price. Considering the disappointment caused by a GB failure, in our basic model, we formulate a two-period game to study the retailer’s optimal pricing decisions and consumers’ purchasing strategies. By deriving the likelihood function of a consumer signing up for the GB deal and utilizing rational expectations theory, we characterize how consumers form their own beliefs about the GB success rate. We find that consumer sentiment toward a failed GB deal serves an important role in a GB deal. Specifically, there exists a threshold of the disappointment factor where the retailer’s profit and consumers’ purchase decisions may change, and consumer leakage and retention may occur. We prove the existence of the subgame perfect Nash equilibrium and outline how the retailer should design the GB schedule. Our study shows that if properly designed, GB is an effective strategy to enhance consumer interests and improve profit. Moreover, a big discount should be offered when the disappointment factor is significant. When the consumer sentiment toward a failed GB deal is insignificant, the retailer should launch a GB deal; otherwise, he should provide the regular sales channel only. We conduct numerical experiments to study the impacts of different factors in a GB deal. Our key results continue to hold in several extensions from our basic model: retailer competition, observable secured group size, and social interactions between consumers.","PeriodicalId":49418,"journal":{"name":"Transportation Research Part E-Logistics and Transportation Review","volume":"63 1","pages":""},"PeriodicalIF":10.6,"publicationDate":"2024-12-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142816529","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 : 2024-12-12DOI: 10.1016/j.tre.2024.103916
Seyed Sina Mohri, Neema Nassir, Russell G. Thompson, Patricia Sauri Lavieri, Hadi Ghaderi
This study designs a crowdshipping (CS) delivery system with public transport (PT) passengers at the operational decision-making level. In this system, parcel lockers (PLs) are positioned in PT stations, through which small and light parcels are allocated to passengers for delivery to their final delivery addresses (i.e., performing the last-mile delivery). A probabilistic mathematical model is formulated with behavioural constraints to estimate the probabilities of accepting CS tasks by passengers. The probability is estimated based on a logit function, sensitive to the parcel’s weight, reimbursement amount, and the walking detour required to deliver the parcel to its final destination. The logit model is constructed based on survey data collected from the Greater Sydney (GS) area, Australia. The mathematical model optimises the allocation of delivery tasks to the CS system and PLs, subsequently, incentivising CS-allocated tasks for participating passengers. Furthermore, the model performs the routing of vehicles to deliver non-allocated parcels, including heavy parcels. A heuristic solution algorithm is then proposed to optimise decisions related to allocation, routing, and incentivisation, which was tested on a real case study. By conducting sensitivity analysis on various model parameters, results show that for a small carrier, utilising a PT-based CS system could minimise daily delivery costs by up to 36%, depending on passengers’ rate of familiarity with the CS initiative and the number of PT stations equipped with PLs. Vehicle delivery cost in the CS-integrated delivery system is also reduced between 50% and 65%, in comparison to the conventional vehicle-only system. Our study reveals that a CS system should offer higher incentives at the beginning, and as CS familiarity grows, figures could be reduced depending on other market and operational conditions. Furthermore, simulated experiments suggest that denser PL networks enable carriers to reduce incentives even at earlier stages with lower familiarity rates.
{"title":"Crowd-shipping systems with public transport passengers: Operational planning","authors":"Seyed Sina Mohri, Neema Nassir, Russell G. Thompson, Patricia Sauri Lavieri, Hadi Ghaderi","doi":"10.1016/j.tre.2024.103916","DOIUrl":"https://doi.org/10.1016/j.tre.2024.103916","url":null,"abstract":"This study designs a crowdshipping (CS) delivery system with public transport (PT) passengers at the operational decision-making level. In this system, parcel lockers (PLs) are positioned in PT stations, through which small and light parcels are allocated to passengers for delivery to their final delivery addresses (i.e., performing the last-mile delivery). A probabilistic mathematical model is formulated with behavioural constraints to estimate the probabilities of accepting CS tasks by passengers. The probability is estimated based on a logit function, sensitive to the parcel’s weight, reimbursement amount, and the walking detour required to deliver the parcel to its final destination. The logit model is constructed based on survey data collected from the Greater Sydney (GS) area, Australia. The mathematical model optimises the allocation of delivery tasks to the CS system and PLs, subsequently, incentivising CS-allocated tasks for participating passengers. Furthermore, the model performs the routing of vehicles to deliver non-allocated parcels, including heavy parcels. A heuristic solution algorithm is then proposed to optimise decisions related to allocation, routing, and incentivisation, which was tested on a real case study. By conducting sensitivity analysis on various model parameters, results show that for a small carrier, utilising a PT-based CS system could minimise daily delivery costs by up to 36%, depending on passengers’ rate of familiarity with the CS initiative and the number of PT stations equipped with PLs. Vehicle delivery cost in the CS-integrated delivery system is also reduced between 50% and 65%, in comparison to the conventional vehicle-only system. Our study reveals that a CS system should offer higher incentives at the beginning, and as CS familiarity grows, figures could be reduced depending on other market and operational conditions. Furthermore, simulated experiments suggest that denser PL networks enable carriers to reduce incentives even at earlier stages with lower familiarity rates.","PeriodicalId":49418,"journal":{"name":"Transportation Research Part E-Logistics and Transportation Review","volume":"119 1","pages":""},"PeriodicalIF":10.6,"publicationDate":"2024-12-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142816527","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}
We study the incentives of an airline and a high-speed rail (HSR) operator to incur sunk costs and cooperate in a hub-and-spoke network with a congested hub airport. Contrary to common wisdom, we find that a high delay cost at the hub reduces incentives to cooperate, and that hub traffic may increase after cooperation. We show that airline-HSR cooperation improves consumer surplus, since higher passenger volumes yield more benefits than incremental delay costs at the hub. We also show that transport operators underinvest in airline-HSR cooperation because (depending on mode substitution and the delay cost) they may not be willing to incur sunk costs when social welfare would be higher under cooperation. We then investigate the rationale and implications of airport price regulation. Finally, we show that transport operators’ and the airport company’s interests may be misaligned, and that airport managers can play a role in encouraging or hindering airline-HSR cooperation, depending on their ability to commit to the airport charge.
{"title":"Airline-High speed rail cooperation, hub congestion, and airport conduct","authors":"Alessandro Avenali, Tiziana D’Alfonso, Pierfrancesco Reverberi","doi":"10.1016/j.tre.2024.103818","DOIUrl":"https://doi.org/10.1016/j.tre.2024.103818","url":null,"abstract":"We study the incentives of an airline and a high-speed rail (HSR) operator to incur sunk costs and cooperate in a hub-and-spoke network with a congested hub airport. Contrary to common wisdom, we find that a high delay cost at the hub reduces incentives to cooperate, and that hub traffic may increase after cooperation. We show that airline-HSR cooperation improves consumer surplus, since higher passenger volumes yield more benefits than incremental delay costs at the hub. We also show that transport operators underinvest in airline-HSR cooperation because (depending on mode substitution and the delay cost) they may not be willing to incur sunk costs when social welfare would be higher under cooperation. We then investigate the rationale and implications of airport price regulation<ce:italic>.</ce:italic> Finally, we show that transport operators’ and the airport company’s interests may be misaligned, and that airport managers can play a role in encouraging or hindering airline-HSR cooperation, depending on their ability to commit to the airport charge.","PeriodicalId":49418,"journal":{"name":"Transportation Research Part E-Logistics and Transportation Review","volume":"44 1","pages":""},"PeriodicalIF":10.6,"publicationDate":"2024-12-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142816531","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 : 2024-12-12DOI: 10.1016/j.tre.2024.103912
Mingyou Meng, Shiming Deng, Pin Zhou, He Xu
To address stakeholders’ interests, firms increasingly adopt a dual-purpose agenda, typically involving the pursuit of profits and consumer surplus (CS). This study considers a supply chain dynamic involving a retailer and a national brand (NB) manufacturer, both potentially pursuing dual purposes, to investigate how their dual-purpose nature influences the introduction and quality strategies of the retailer’s store brand (SB). Our findings show that only the retailer pursuing CS has no impact on SB quality. However, if the NB manufacturer pursues CS, SB quality declines irrespective of the retailer’s stance. Interestingly, the for-profit retailer experiences reduced profits from SB introduction when also pursuing CS and expressing a high interest in it. Conversely, SB introduction enhances the dual-purpose manufacturer’s utility when its interest in CS is relatively high. The introduction of SB may lead to unintended price and payoff implications, with the manufacturer’s profit and the wholesale price exhibiting non-monotonic relationships with its interest in CS. Consequently, compared to the for-profit scenario, this may elevate the wholesale price, exacerbating the double marginalization effect. Additionally, when the retailer pursues CS, supply chain profit may increase because of the mitigated double marginalization effect, resulting from an unconventional reduction in retail markup rather than wholesale price. Our findings suggest that manufacturers pursuing CS could strategically alleviate profit losses stemming from retailers’ SB introduction. However, retailers should exercise caution when simultaneously introducing a SB and pursuing CS from a profitability standpoint.
{"title":"The role of dual purpose in retailer’s store brand introduction and quality strategies within a supply chain","authors":"Mingyou Meng, Shiming Deng, Pin Zhou, He Xu","doi":"10.1016/j.tre.2024.103912","DOIUrl":"https://doi.org/10.1016/j.tre.2024.103912","url":null,"abstract":"To address stakeholders’ interests, firms increasingly adopt a dual-purpose agenda, typically involving the pursuit of profits and consumer surplus (CS). This study considers a supply chain dynamic involving a retailer and a national brand (NB) manufacturer, both potentially pursuing dual purposes, to investigate how their dual-purpose nature influences the introduction and quality strategies of the retailer’s store brand (SB). Our findings show that only the retailer pursuing CS has no impact on SB quality. However, if the NB manufacturer pursues CS, SB quality declines irrespective of the retailer’s stance. Interestingly, the for-profit retailer experiences reduced profits from SB introduction when also pursuing CS and expressing a high interest in it. Conversely, SB introduction enhances the dual-purpose manufacturer’s utility when its interest in CS is relatively high. The introduction of SB may lead to unintended price and payoff implications, with the manufacturer’s profit and the wholesale price exhibiting non-monotonic relationships with its interest in CS. Consequently, compared to the for-profit scenario, this may elevate the wholesale price, exacerbating the double marginalization effect. Additionally, when the retailer pursues CS, supply chain profit may increase because of the mitigated double marginalization effect, resulting from an unconventional reduction in retail markup rather than wholesale price. Our findings suggest that manufacturers pursuing CS could strategically alleviate profit losses stemming from retailers’ SB introduction. However, retailers should exercise caution when simultaneously introducing a SB and pursuing CS from a profitability standpoint.","PeriodicalId":49418,"journal":{"name":"Transportation Research Part E-Logistics and Transportation Review","volume":"8 1","pages":""},"PeriodicalIF":10.6,"publicationDate":"2024-12-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142816526","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 : 2024-12-11DOI: 10.1016/j.tre.2024.103900
Tao Tang, Simin Chai, Wei Wu, Jiateng Yin, Andrea D’Ariano
In high-speed railway systems, unexpected disruptions can result in delays of trains, significantly affecting the quality of service for passengers. Train Timetable Rescheduling (TTR) is a crucial task in the daily operation of high-speed railways to maintain punctuality and efficiency in the face of such unforeseen disruptions. Most existing studies on TTR are based on integer programming (IP) techniques and are required to solve IP models repetitively in case of disruptions, which however may be very time-consuming and greatly limit their usefulness in practice. Our study first proposes a multi-task deep reinforcement learning (MDRL) approach for TTR. Our MDRL is constructed and trained offline with a large number of historical disruptive events, enabling to generate TTR decisions in real-time for different disruption cases. Specifically, we transform the TTR problem into a Markov decision process considering the retiming and rerouting of trains. Then, we construct the MDRL framework with the definition of state, action, transition, reward, and value function approximations with neural networks for each agent (i.e., rail train), by considering the information of different disruption events as tasks. To overcome the low training efficiency and huge memory usage in the training of MDRL, given a large number of disruptive events in the historical data, we develop a new and high-efficient training method based on a Quadratic assignment programming (QAP) model and a Frank-Wolfe-based algorithm. Our QAP model optimizes only a small number but most “representative” tasks from the historical data, while the Frank-Wolfe-based algorithm approximates the nonlinear terms in the value function of MDRL and updates the model parameters among different training tasks concurrently. Finally, based on the real-world data from the Beijing–Zhangjiakou high-speed railway systems, we evaluate the performance of our MDRL approach by benchmarking it against state-of-the-art approaches in the literature. Our computational results demonstrate that an offline-trained MDRL is able to generate near-optimal TTR solutions in real-time against different disruption scenarios, and it evidently outperforms state-of-art models regarding solution quality and computational time.
{"title":"A multi-task deep reinforcement learning approach to real-time railway train rescheduling","authors":"Tao Tang, Simin Chai, Wei Wu, Jiateng Yin, Andrea D’Ariano","doi":"10.1016/j.tre.2024.103900","DOIUrl":"https://doi.org/10.1016/j.tre.2024.103900","url":null,"abstract":"In high-speed railway systems, unexpected disruptions can result in delays of trains, significantly affecting the quality of service for passengers. Train Timetable Rescheduling (TTR) is a crucial task in the daily operation of high-speed railways to maintain punctuality and efficiency in the face of such unforeseen disruptions. Most existing studies on TTR are based on integer programming (IP) techniques and are required to solve IP models repetitively in case of disruptions, which however may be very time-consuming and greatly limit their usefulness in practice. Our study first proposes a multi-task deep reinforcement learning (MDRL) approach for TTR. Our MDRL is constructed and trained offline with a large number of historical disruptive events, enabling to generate TTR decisions in real-time for different disruption cases. Specifically, we transform the TTR problem into a Markov decision process considering the retiming and rerouting of trains. Then, we construct the MDRL framework with the definition of state, action, transition, reward, and value function approximations with neural networks for each agent (i.e., rail train), by considering the information of different disruption events as tasks. To overcome the low training efficiency and huge memory usage in the training of MDRL, given a large number of disruptive events in the historical data, we develop a new and high-efficient training method based on a Quadratic assignment programming (QAP) model and a Frank-Wolfe-based algorithm. Our QAP model optimizes only a small number but most “representative” tasks from the historical data, while the Frank-Wolfe-based algorithm approximates the nonlinear terms in the value function of MDRL and updates the model parameters among different training tasks concurrently. Finally, based on the real-world data from the Beijing–Zhangjiakou high-speed railway systems, we evaluate the performance of our MDRL approach by benchmarking it against state-of-the-art approaches in the literature. Our computational results demonstrate that an offline-trained MDRL is able to generate near-optimal TTR solutions in real-time against different disruption scenarios, and it evidently outperforms state-of-art models regarding solution quality and computational time.","PeriodicalId":49418,"journal":{"name":"Transportation Research Part E-Logistics and Transportation Review","volume":"50 1","pages":""},"PeriodicalIF":10.6,"publicationDate":"2024-12-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142816530","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 : 2024-12-11DOI: 10.1016/j.tre.2024.103898
Yilang Hao, Zhibin Chen, Xiaotong Sun, Lu Tong
Truck platooning, a linking technology of trucks on the highway, has gained enormous attention in recent years due to its benefits in energy and operation cost savings. However, most existing studies on truck platooning limit their focus on particular scenarios that each truck can serve only one customer demand and is thus with a specified origin–destination pair, so only routing and time schedules are taken into account. Nevertheless, in real-world logistics, each truck may need to serve multiple customers located at different places, and the operator managing a fleet of trucks thus has to determine not only the routing and time schedules of each truck but also the set of customers allocated to each truck and their sequence to visit. This is well known as a capacitated vehicle routing problem with time windows (CVRPTW), and considering the application of truck platooning in such a problem entails new modeling frameworks and tailored solution algorithms. In light of this, this study makes the first attempt to optimize the truck platooning plan for a road-network CVRPTW in a way to minimize the total operation cost, including vehicles’ fixed dispatch cost and energy cost, while fulfilling all delivery demands within their time window constraints. Specifically, the operation plan will dictate the number of trucks to be dispatched, the set of customers, and the routing and time schedules for each truck. In addition, the modeling framework is constructed based on a road network instead of a traditional customer node graph to better resemble and facilitate the platooning operation. A 3-stage algorithm embedded with a ”route-then-schedule” scheme, Dynamic Programming, and Modified Insertion heuristic, is developed to solve the proposed model in a timely manner. Numerical experiments are conducted to validate the proposed modeling framework, demonstrate the performance of the proposed solution algorithm, and quantify the benefit brought by the truck platooning technology.
{"title":"Planning of truck platooning for road-network capacitated vehicle routing problem","authors":"Yilang Hao, Zhibin Chen, Xiaotong Sun, Lu Tong","doi":"10.1016/j.tre.2024.103898","DOIUrl":"https://doi.org/10.1016/j.tre.2024.103898","url":null,"abstract":"Truck platooning, a linking technology of trucks on the highway, has gained enormous attention in recent years due to its benefits in energy and operation cost savings. However, most existing studies on truck platooning limit their focus on particular scenarios that each truck can serve only one customer demand and is thus with a specified origin–destination pair, so only routing and time schedules are taken into account. Nevertheless, in real-world logistics, each truck may need to serve multiple customers located at different places, and the operator managing a fleet of trucks thus has to determine not only the routing and time schedules of each truck but also the set of customers allocated to each truck and their sequence to visit. This is well known as a capacitated vehicle routing problem with time windows (CVRPTW), and considering the application of truck platooning in such a problem entails new modeling frameworks and tailored solution algorithms. In light of this, this study makes the first attempt to optimize the truck platooning plan for a road-network CVRPTW in a way to minimize the total operation cost, including vehicles’ fixed dispatch cost and energy cost, while fulfilling all delivery demands within their time window constraints. Specifically, the operation plan will dictate the number of trucks to be dispatched, the set of customers, and the routing and time schedules for each truck. In addition, the modeling framework is constructed based on a road network instead of a traditional customer node graph to better resemble and facilitate the platooning operation. A 3-stage algorithm embedded with a ”route-then-schedule” scheme, Dynamic Programming, and Modified Insertion heuristic, is developed to solve the proposed model in a timely manner. Numerical experiments are conducted to validate the proposed modeling framework, demonstrate the performance of the proposed solution algorithm, and quantify the benefit brought by the truck platooning technology.","PeriodicalId":49418,"journal":{"name":"Transportation Research Part E-Logistics and Transportation Review","volume":"5 1","pages":""},"PeriodicalIF":10.6,"publicationDate":"2024-12-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142816532","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 : 2024-12-09DOI: 10.1016/j.tre.2024.103883
Zihan Qiu, Jiancheng Long, Yang Yu, Shukai Chen
This paper considers an intelligent warehouse system (IWS) that requires the seamless cooperation of three types of mobile robots: automated guided vehicles (AGVs), rail-guided vehicles (RGVs), and gantry lifting devices (GLDs). Compared to the conventional system, which comprises AGVs, the IWS is more flexible in addressing with the customized demands of diverse enterprises. This paper proposes an integrated task assignment and path planning problem for multi-type robots (e.g., AGVs, RGVs, and GLDs) in IWS. The cooperative constraints between AGVs and GLDs, RGVs and GLDs, as well as the conflict-free constraints among AGVs, are considered. It is challenging to solve the multi-type robots scheduling problem with the conflict-free constraints of AGVs because these constraints can result in the unfixed task completion time of AGVs and pose computational challenges of the task assignment for AGVs, RGVs, and GLDs. The proposed integrated task assignment and path planning problem for multi-type robots is modeled as a multi-commodity flow problem on a novel state-time–space network and is formulated as an integer linear programming (ILP) model, where the warehouse operator aims to minimize the total completion time of all tasks. We developed a Lagrangian relaxation heuristic with a customized efficient strategy to find feasible solutions. We also solved our proposed model using CPLEX. The tailored Lagrangian relaxation heuristic was tested on simulated and real instances provided by a manufacturing company. The results show that the proposed heuristic outperforms the baseline algorithm. Sensitivity analyses from the numerical experiments are discussed, which can help the company improve the efficiency of the IWS.
{"title":"Integrated task assignment and path planning for multi-type robots in an intelligent warehouse system","authors":"Zihan Qiu, Jiancheng Long, Yang Yu, Shukai Chen","doi":"10.1016/j.tre.2024.103883","DOIUrl":"https://doi.org/10.1016/j.tre.2024.103883","url":null,"abstract":"This paper considers an intelligent warehouse system (IWS) that requires the seamless cooperation of three types of mobile robots: automated guided vehicles (AGVs), rail-guided vehicles (RGVs), and gantry lifting devices (GLDs). Compared to the conventional system, which comprises AGVs, the IWS is more flexible in addressing with the customized demands of diverse enterprises. This paper proposes an integrated task assignment and path planning problem for multi-type robots (e.g., AGVs, RGVs, and GLDs) in IWS. The cooperative constraints between AGVs and GLDs, RGVs and GLDs, as well as the conflict-free constraints among AGVs, are considered. It is challenging to solve the multi-type robots scheduling problem with the conflict-free constraints of AGVs because these constraints can result in the unfixed task completion time of AGVs and pose computational challenges of the task assignment for AGVs, RGVs, and GLDs. The proposed integrated task assignment and path planning problem for multi-type robots is modeled as a multi-commodity flow problem on a novel state-time–space network and is formulated as an integer linear programming (ILP) model, where the warehouse operator aims to minimize the total completion time of all tasks. We developed a Lagrangian relaxation heuristic with a customized efficient strategy to find feasible solutions. We also solved our proposed model using CPLEX. The tailored Lagrangian relaxation heuristic was tested on simulated and real instances provided by a manufacturing company. The results show that the proposed heuristic outperforms the baseline algorithm. Sensitivity analyses from the numerical experiments are discussed, which can help the company improve the efficiency of the IWS.","PeriodicalId":49418,"journal":{"name":"Transportation Research Part E-Logistics and Transportation Review","volume":"21 1","pages":""},"PeriodicalIF":10.6,"publicationDate":"2024-12-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142816536","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 : 2024-12-09DOI: 10.1016/j.tre.2024.103899
Xin Chen, Gege Jiang, Yu Jiang
The recent application of the rational inattention (RI) theory in transportation has shed light on a promising alternative way of understanding how information influences the travel choices of passengers. However, existing RI literature has not yet addressed the discrete choice problem with multiple variates. Thus, this study develops a multivariate rational inattention (MRI) discrete choice model. This assumes that acquiring information is costly and the unit information cost varies among variates, so decision-makers rationally choose the amount of information to acquire for each variate. We demonstrate that the MRI discrete choice model results in a probabilistic formulation similar to the logit model, but with the superiority of integrating unit information costs and the prior knowledge of decision-makers. Furthermore, we apply the MRI discrete choice model to the metro route choice problem and calibrate the model based on the revealed preference (RP) data collected from the Chengdu metro. It is found that the proposed model has satisfactory accuracy with better interpretability than the logit model and univariate rational inattention discrete choice model.
{"title":"Multivariate discrete choice with rational inattention: Model development, application, and calibration","authors":"Xin Chen, Gege Jiang, Yu Jiang","doi":"10.1016/j.tre.2024.103899","DOIUrl":"https://doi.org/10.1016/j.tre.2024.103899","url":null,"abstract":"The recent application of the rational inattention (RI) theory in transportation has shed light on a promising alternative way of understanding how information influences the travel choices of passengers. However, existing RI literature has not yet addressed the discrete choice problem with multiple variates. Thus, this study develops a multivariate rational inattention (MRI) discrete choice model. This assumes that acquiring information is costly and the unit information cost varies among variates, so decision-makers rationally choose the amount of information to acquire for each variate. We demonstrate that the MRI discrete choice model results in a probabilistic formulation similar to the logit model, but with the superiority of integrating unit information costs and the prior knowledge of decision-makers. Furthermore, we apply the MRI discrete choice model to the metro route choice problem and calibrate the model based on the revealed preference (RP) data collected from the Chengdu metro. It is found that the proposed model has satisfactory accuracy with better interpretability than the logit model and univariate rational inattention discrete choice model.","PeriodicalId":49418,"journal":{"name":"Transportation Research Part E-Logistics and Transportation Review","volume":"53 4 1","pages":""},"PeriodicalIF":10.6,"publicationDate":"2024-12-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142816533","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}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：