Pub Date : 2026-04-16Epub Date: 2025-08-28DOI: 10.1016/j.ejor.2025.08.040
F. Cesarone, M. Corradini, L. Lampariello, J. Riccioni
We focus on a behavioral model that has been recently proposed in the literature, whose rationale can be traced back to the Half-Full/Half-Empty glass metaphor. More precisely, we generalize the Half-Full/Half-Empty approach to the context of positive and negative lotteries and give financial and behavioral interpretations of the Half-Full/Half-Empty parameters. We develop a portfolio selection model based on the Half-Full/Half-Empty strategy, resulting in a nonconvex optimization problem, which, nonetheless, is proven to be equivalent to an alternative Mixed-Integer Linear Programming formulation. Based on three real-world datasets, we obtain empirical validation of the theoretical properties of the Half-Full/Half-Empty model, and the computational results highlight the versatility of our approach when varying its defining parameter values.
{"title":"A new behavioral model for portfolio selection using the Half-Full/Half-Empty approach","authors":"F. Cesarone, M. Corradini, L. Lampariello, J. Riccioni","doi":"10.1016/j.ejor.2025.08.040","DOIUrl":"10.1016/j.ejor.2025.08.040","url":null,"abstract":"<div><div>We focus on a behavioral model that has been recently proposed in the literature, whose rationale can be traced back to the Half-Full/Half-Empty glass metaphor. More precisely, we generalize the Half-Full/Half-Empty approach to the context of positive and negative lotteries and give financial and behavioral interpretations of the Half-Full/Half-Empty parameters. We develop a portfolio selection model based on the Half-Full/Half-Empty strategy, resulting in a nonconvex optimization problem, which, nonetheless, is proven to be equivalent to an alternative Mixed-Integer Linear Programming formulation. Based on three real-world datasets, we obtain empirical validation of the theoretical properties of the Half-Full/Half-Empty model, and the computational results highlight the versatility of our approach when varying its defining parameter values.</div></div>","PeriodicalId":55161,"journal":{"name":"European Journal of Operational Research","volume":"330 2","pages":"Pages 687-699"},"PeriodicalIF":6.0,"publicationDate":"2026-04-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144923967","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"管理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-04-16Epub Date: 2025-09-11DOI: 10.1016/j.ejor.2025.08.010
Yihua Wang , Stefan Minner
In a two-echelon supply chain, achieving high service levels at minimal cost can necessitate the use of flexible supply options, such as lateral transshipment and emergency delivery, particularly during stockouts at local warehouses. However, these flexible supply options are only beneficial if they enable deliveries to arrive earlier than regular replenishment orders. Therefore, it is essential to track the delivery of outstanding orders and obtain information about the supply pipeline, especially for spare parts networks where demand is typically slow-moving. Advances in real-time tracking technologies, such as the Internet of Things, provide critical visibility into the supply pipeline. We propose a flexible supply strategy that incorporates real-time pipeline information with regard to the position of regular replenishment orders. We develop a model to analyze the fraction of each supply option chosen and the long-term average operational cost under the proposed flexible supply strategy. By comparing the proposed model with a baseline model that excludes pipeline information, our numerical study reveals an average cost saving of 5.7% from using pipeline information. Furthermore, incorporating pipeline information impacts optimal inventory allocation, reducing the target inventory level at the central warehouse and shifting more inventory downstream to local warehouses. Pipeline visibility reduces the reliance on expensive emergency deliveries, leading to a more sustainable and agile supply chain.
{"title":"Supply flexibility in two-echelon stochastic spare parts inventory systems with real-time pipeline information","authors":"Yihua Wang , Stefan Minner","doi":"10.1016/j.ejor.2025.08.010","DOIUrl":"10.1016/j.ejor.2025.08.010","url":null,"abstract":"<div><div>In a two-echelon supply chain, achieving high service levels at minimal cost can necessitate the use of flexible supply options, such as lateral transshipment and emergency delivery, particularly during stockouts at local warehouses. However, these flexible supply options are only beneficial if they enable deliveries to arrive earlier than regular replenishment orders. Therefore, it is essential to track the delivery of outstanding orders and obtain information about the supply pipeline, especially for spare parts networks where demand is typically slow-moving. Advances in real-time tracking technologies, such as the Internet of Things, provide critical visibility into the supply pipeline. We propose a flexible supply strategy that incorporates real-time pipeline information with regard to the position of regular replenishment orders. We develop a model to analyze the fraction of each supply option chosen and the long-term average operational cost under the proposed flexible supply strategy. By comparing the proposed model with a baseline model that excludes pipeline information, our numerical study reveals an average cost saving of 5.7% from using pipeline information. Furthermore, incorporating pipeline information impacts optimal inventory allocation, reducing the target inventory level at the central warehouse and shifting more inventory downstream to local warehouses. Pipeline visibility reduces the reliance on expensive emergency deliveries, leading to a more sustainable and agile supply chain.</div></div>","PeriodicalId":55161,"journal":{"name":"European Journal of Operational Research","volume":"330 2","pages":"Pages 458-472"},"PeriodicalIF":6.0,"publicationDate":"2026-04-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145094157","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"管理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-04-16Epub Date: 2025-09-26DOI: 10.1016/j.ejor.2025.09.026
Khadija Hadj Salem , Arthur Kramer , Alexis Robbes
The job sequencing and tool switching problem with non-identical parallel machines (SSP-NPM) is a generalization of the job sequencing and tool switching problem (SSP), which is known to be an -hard optimization problem. SSP-NPM involves scheduling jobs on non-identical parallel machines while determining the associated tool sequences to minimize the makespan. In this paper, we propose an improved version of the existing position-based MILP model introduced in Calmels (2022a). Building on this, we develop a position-based arc flow model. In addition, we introduce a novel MILP formulation based on a job group representation of the problem, as used in Akhundov and Ostrowski (2024) for the SSP, which we extend to derive a job group based arc flow model. Several valid inequalities and symmetry-breaking constraints were considered to improve the proposed approaches. Lower and upper bounds were also developed to improve the proposed approaches. Computational experiments were performed using available literature and randomly generated instances to evaluate the effectiveness of the proposed approaches.
{"title":"Job sequencing and tool switching problem with non-identical parallel machines: mathematical formulations and modeling improvements","authors":"Khadija Hadj Salem , Arthur Kramer , Alexis Robbes","doi":"10.1016/j.ejor.2025.09.026","DOIUrl":"10.1016/j.ejor.2025.09.026","url":null,"abstract":"<div><div>The job sequencing and tool switching problem with non-identical parallel machines (SSP-NPM) is a generalization of the job sequencing and tool switching problem (SSP), which is known to be an <span><math><mi>NP</mi></math></span>-hard optimization problem. SSP-NPM involves scheduling jobs on non-identical parallel machines while determining the associated tool sequences to minimize the makespan. In this paper, we propose an improved version of the existing position-based MILP model introduced in Calmels (2022a). Building on this, we develop a position-based arc flow model. In addition, we introduce a novel MILP formulation based on a job group representation of the problem, as used in Akhundov and Ostrowski (2024) for the SSP, which we extend to derive a job group based arc flow model. Several valid inequalities and symmetry-breaking constraints were considered to improve the proposed approaches. Lower and upper bounds were also developed to improve the proposed approaches. Computational experiments were performed using available literature and randomly generated instances to evaluate the effectiveness of the proposed approaches.</div></div>","PeriodicalId":55161,"journal":{"name":"European Journal of Operational Research","volume":"330 2","pages":"Pages 416-426"},"PeriodicalIF":6.0,"publicationDate":"2026-04-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145228835","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"管理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-04-16Epub Date: 2025-08-15DOI: 10.1016/j.ejor.2025.08.022
Ruxiao Xing, Bo Li
With the rapid development of technology, big data-driven information services are gradually becoming an important tool for firms on e-commerce platforms. Firms can achieve more accurate demand forecasting and optimize the decision-making process through information. This paper examines how big-data-driven information services affect supply chain decisions by constructing Stackelberg models involving a platform and a manufacturer under two cooperation modes: reselling and agency. The platform determines the skill level of data analytics and provides two information services with different data amount, whereas the manufacturer decides which one to purchase. These choices jointly affect the forecasting accuracy for demand. Through comparison analyses, we find that information services exhibit a significant leverage effect, which confirms the value of information. However, it is unexpected that information services are not always effective in counteracting demand uncertainty, which is related to the investment cost of data analytics skill level. Additionally, the efficiency of information services also differs under two modes. With high investment cost, information services are efficient in reselling mode but inefficient in agency mode. Finally, we reveal that the manufacturer favours reselling mode when the information service fee is low and investment cost is moderate. Otherwise, he selects agency mode. The platform prefers agency mode only when the information service fee is high and investment cost is low. Both parties can achieve win-win situations under certain conditions. These findings clarify the value of information services under different cooperation modes, providing a key reference for cooperation between the platform and manufacturer.
{"title":"The effects of big-data-driven information service sales on cooperation modes in a supply chain","authors":"Ruxiao Xing, Bo Li","doi":"10.1016/j.ejor.2025.08.022","DOIUrl":"10.1016/j.ejor.2025.08.022","url":null,"abstract":"<div><div>With the rapid development of technology, big data-driven information services are gradually becoming an important tool for firms on e-commerce platforms. Firms can achieve more accurate demand forecasting and optimize the decision-making process through information. This paper examines how big-data-driven information services affect supply chain decisions by constructing Stackelberg models involving a platform and a manufacturer under two cooperation modes: reselling and agency. The platform determines the skill level of data analytics and provides two information services with different data amount, whereas the manufacturer decides which one to purchase. These choices jointly affect the forecasting accuracy for demand. Through comparison analyses, we find that information services exhibit a significant leverage effect, which confirms the value of information. However, it is unexpected that information services are not always effective in counteracting demand uncertainty, which is related to the investment cost of data analytics skill level. Additionally, the efficiency of information services also differs under two modes. With high investment cost, information services are efficient in reselling mode but inefficient in agency mode. Finally, we reveal that the manufacturer favours reselling mode when the information service fee is low and investment cost is moderate. Otherwise, he selects agency mode. The platform prefers agency mode only when the information service fee is high and investment cost is low. Both parties can achieve win-win situations under certain conditions. These findings clarify the value of information services under different cooperation modes, providing a key reference for cooperation between the platform and manufacturer.</div></div>","PeriodicalId":55161,"journal":{"name":"European Journal of Operational Research","volume":"330 2","pages":"Pages 640-655"},"PeriodicalIF":6.0,"publicationDate":"2026-04-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144900171","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"管理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-04-16Epub Date: 2025-08-08DOI: 10.1016/j.ejor.2025.07.040
Alexander Bloemer , Stefan Minner
Governments and consumers increasingly hold companies responsible for sustainability violations across their supply chain. However, companies often have only limited information on sustainability practices in their upstream supply chain. To still achieve sustainable behavior of lower-tier suppliers, they have to choose adequate incentives or delegate this responsibility to their direct suppliers. We analyze a three-tier supply chain where a buyer and a direct supplier can motivate sustainable behavior of a sub-supplier by auditing or investing in training.
Auditing and training are substitutable for the buyer in a two-tier supply chain but can become complementary when being able to delegate the incentive effort to the direct supplier. As the external investigation level increases, the buyer and the direct supplier do not necessarily switch to training, but may rely more on auditing. Such increasing external investigations, just as higher penalties, can reduce the sustainability effort through the option of delegating the incentive effort to the direct supplier. Likewise, combining both mechanisms can also incentivize a lower sustainability effort.
{"title":"Auditing and training to incentivize sustainability in multi-tier supply chains: Substitutes or complements?","authors":"Alexander Bloemer , Stefan Minner","doi":"10.1016/j.ejor.2025.07.040","DOIUrl":"10.1016/j.ejor.2025.07.040","url":null,"abstract":"<div><div>Governments and consumers increasingly hold companies responsible for sustainability violations across their supply chain. However, companies often have only limited information on sustainability practices in their upstream supply chain. To still achieve sustainable behavior of lower-tier suppliers, they have to choose adequate incentives or delegate this responsibility to their direct suppliers. We analyze a three-tier supply chain where a buyer and a direct supplier can motivate sustainable behavior of a sub-supplier by auditing or investing in training.</div><div>Auditing and training are substitutable for the buyer in a two-tier supply chain but can become complementary when being able to delegate the incentive effort to the direct supplier. As the external investigation level increases, the buyer and the direct supplier do not necessarily switch to training, but may rely more on auditing. Such increasing external investigations, just as higher penalties, can reduce the sustainability effort through the option of delegating the incentive effort to the direct supplier. Likewise, combining both mechanisms can also incentivize a lower sustainability effort.</div></div>","PeriodicalId":55161,"journal":{"name":"European Journal of Operational Research","volume":"330 2","pages":"Pages 656-670"},"PeriodicalIF":6.0,"publicationDate":"2026-04-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144900320","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"管理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Consumer returns are common in live streaming selling, and metaverse technology is gradually advocated to solve the issue of consumer returns since this technology can significantly enhance the live streaming shopping experience. We consider a supply chain composed of a manufacturer and a retailer with two live streaming scenarios: ML and RL. In scenario ML (RL), the manufacturer (retailer) invites an influencer to open a live streaming channel. We explore the impacts of influencer power and consider the integration of metaverse technology. Our analytical results reveal the following: Under scenario ML, the adoption of metaverse technology increases the manufacturer’s optimal profit at a moderate unit cost of metaverse technology with high influencer power or at a low unit cost. When the manufacturers introduce live streaming, they should adopt metaverse technology when choosing influencers with relatively high influencer power as the unit cost is moderate in the development phase of metaverse technology. Under scenario RL, the retailer should only adopt metaverse technology at a high unit cost with low influencer power or at a low unit cost with high influencer power. This implies a counterintuitive strategy: when cooperating with influencers with lower influencer power, the retailers should adopt metaverse technology despite high costs. Furthermore, regardless of metaverse technology adoption, scenario RL (ML) brings more profit to the manufacturer if influencer power is low (high). This indicates that manufacturers and retailers should choose influencers with greater influence power when introducing live streaming. By extending our model, we examine the robustness of our findings under the scenarios involving channel competition, omni-channel strategy, and live streaming slicing. This study is the first to explore the adoption strategies of metaverse technology in different live streaming supply chain structures considering consumer returns, providing valuable guidance for manufacturers and retailers on when to adopt metaverse technology.
{"title":"When should metaverse technology be adopted in live streaming considering consumer returns?","authors":"Xiaoping Xu , Dianming Chen , T.C.E. Cheng , Tsan-Ming Choi","doi":"10.1016/j.ejor.2025.08.019","DOIUrl":"10.1016/j.ejor.2025.08.019","url":null,"abstract":"<div><div>Consumer returns are common in live streaming selling, and metaverse technology is gradually advocated to solve the issue of consumer returns since this technology can significantly enhance the live streaming shopping experience. We consider a supply chain composed of a manufacturer and a retailer with two live streaming scenarios: ML and RL. In scenario ML (RL), the manufacturer (retailer) invites an influencer to open a live streaming channel. We explore the impacts of influencer power and consider the integration of metaverse technology. Our analytical results reveal the following: Under scenario ML, the adoption of metaverse technology increases the manufacturer’s optimal profit at a moderate unit cost of metaverse technology with high influencer power or at a low unit cost. When the manufacturers introduce live streaming, they should adopt metaverse technology when choosing influencers with relatively high influencer power as the unit cost is moderate in the development phase of metaverse technology. Under scenario RL, the retailer should only adopt metaverse technology at a high unit cost with low influencer power or at a low unit cost with high influencer power. This implies a counterintuitive strategy: when cooperating with influencers with lower influencer power, the retailers should adopt metaverse technology despite high costs. Furthermore, regardless of metaverse technology adoption, scenario RL (ML) brings more profit to the manufacturer if influencer power is low (high). This indicates that manufacturers and retailers should choose influencers with greater influence power when introducing live streaming. By extending our model, we examine the robustness of our findings under the scenarios involving channel competition, omni-channel strategy, and live streaming slicing. This study is the first to explore the adoption strategies of metaverse technology in different live streaming supply chain structures considering consumer returns, providing valuable guidance for manufacturers and retailers on when to adopt metaverse technology.</div></div>","PeriodicalId":55161,"journal":{"name":"European Journal of Operational Research","volume":"330 2","pages":"Pages 487-511"},"PeriodicalIF":6.0,"publicationDate":"2026-04-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144901756","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"管理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-04-16Epub Date: 2025-06-26DOI: 10.1016/j.ejor.2025.06.014
C. Archetti , L.C. Coelho , M.G. Speranza , P. Vansteenwegen
Vehicle routing problems (VRPs) are a large class of well-studied and computationally hard combinatorial optimization problems. In the classical capacitated VRP, a fleet of homogeneous capacitated vehicles, that start and end their route at a depot, must satisfy customers’ demand at minimum cost. Many variants and extensions of this problem have been studied. In this paper, after discussing the capacitated VRP, we review the literature for the most-studied variants and extensions of the VRP and then focus on the most recent developments and trends.
{"title":"Beyond fifty years of vehicle routing: Insights into the history and the future","authors":"C. Archetti , L.C. Coelho , M.G. Speranza , P. Vansteenwegen","doi":"10.1016/j.ejor.2025.06.014","DOIUrl":"10.1016/j.ejor.2025.06.014","url":null,"abstract":"<div><div>Vehicle routing problems (VRPs) are a large class of well-studied and computationally hard combinatorial optimization problems. In the classical capacitated VRP, a fleet of homogeneous capacitated vehicles, that start and end their route at a depot, must satisfy customers’ demand at minimum cost. Many variants and extensions of this problem have been studied. In this paper, after discussing the capacitated VRP, we review the literature for the most-studied variants and extensions of the VRP and then focus on the most recent developments and trends.</div></div>","PeriodicalId":55161,"journal":{"name":"European Journal of Operational Research","volume":"330 2","pages":"Pages 355-372"},"PeriodicalIF":6.0,"publicationDate":"2026-04-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144515810","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"管理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-04-16Epub Date: 2025-08-27DOI: 10.1016/j.ejor.2025.08.045
Baozhuang Niu, Jiayun Liu, Yiyuan Ruan
In the cross-border e-commerce era, many multinational firms (MNFs) are operating local retailing divisions in emerging markets while also selling through e-tailers. Though e-tailers usually have a demand information advantage that can be known by the MNFs’ manufacturing divisions, interestingly, we note that due to a lack of cyber resilience, data transfer barriers between MNFs’ manufacturing and retailing divisions have widely inhibited intersectoral information transfer. Therefore, risk-averse retailing divisions may postpone orders to wait for the information, exposing them to a second-mover disadvantage. Intuitively, advanced intersectoral data transfer such as AI-enabled cybersecurity cloud platform is capable of facilitating the retailing division’s timely order based on demand information shared by the MNF, we observe that MNF’s tax-planning may be hampered under the Arm’s Length Principle, and tensions between horizontal (i.e., the e-tailer and the MNF’s retailing division) and vertical coordination may be exacerbated. Therefore, we study the MNF’s incentives to apply intersectoral data transfer based on a co-opetition game-theoretic model. We show that the retailing division’s preference varies with the tax disparity and its risk-averse degree. When tax disparity narrows, the sourcing cost-saving effect from postponed ordering allows the division to access demand information without incurring high costs. Establishing intersectoral data transfer is not always optimal for either the MNF or its retailing division. The MNF benefits when the tax disparity is high, whereas the retailing division should balance the information value against sourcing cost saving when deciding order timing.
{"title":"Should a multinational firm incentivize timely orders from retailing division through intersectoral data transfer?","authors":"Baozhuang Niu, Jiayun Liu, Yiyuan Ruan","doi":"10.1016/j.ejor.2025.08.045","DOIUrl":"10.1016/j.ejor.2025.08.045","url":null,"abstract":"<div><div>In the cross-border e-commerce era, many multinational firms (MNFs) are operating local retailing divisions in emerging markets while also selling through e-tailers. Though e-tailers usually have a demand information advantage that can be known by the MNFs’ manufacturing divisions, interestingly, we note that due to a lack of cyber resilience, data transfer barriers between MNFs’ manufacturing and retailing divisions have widely inhibited intersectoral information transfer. Therefore, risk-averse retailing divisions may postpone orders to wait for the information, exposing them to a second-mover disadvantage. Intuitively, advanced intersectoral data transfer such as AI-enabled cybersecurity cloud platform is capable of facilitating the retailing division’s timely order based on demand information shared by the MNF, we observe that MNF’s tax-planning may be hampered under the Arm’s Length Principle, and tensions between horizontal (i.e., the e-tailer and the MNF’s retailing division) and vertical coordination may be exacerbated. Therefore, we study the MNF’s incentives to apply intersectoral data transfer based on a co-opetition game-theoretic model. We show that the retailing division’s preference varies with the tax disparity and its risk-averse degree. When tax disparity narrows, the <em>sourcing cost-saving effect</em> from postponed ordering allows the division to access demand information without incurring high costs. Establishing intersectoral data transfer is not always optimal for either the MNF or its retailing division. The MNF benefits when the tax disparity is high, whereas the retailing division should balance the information value against sourcing cost saving when deciding order timing.</div></div>","PeriodicalId":55161,"journal":{"name":"European Journal of Operational Research","volume":"330 2","pages":"Pages 627-639"},"PeriodicalIF":6.0,"publicationDate":"2026-04-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145009230","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"管理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-04-16Epub Date: 2025-12-06DOI: 10.1016/j.ejor.2025.12.001
Donatien Hainaut , Jean-Loup Dupret
This article introduces a novel iterative algorithm that combines policy iterations with Gaussian process regressions to solve stochastic control problems. At each iteration, an approximate value function is updated, followed by an improvement of the control strategy. State variables are sampled both within the interior domain and on the terminal boundary of the Hamilton-Jacobi-Bellman (HJB) equation. The approximate value function, interpreted as the solution to the HJB equation, is obtained by fitting a constrained regression model. This regression function aligns with the terminal utility on the boundary sample and satisfies the HJB equation on the interior sample. Assuming the regression function follows a Gaussian process, we derive closed-form approximations for both the value function and its derivatives. A numerical illustration demonstrates the efficiency of the proposed method in solving the consumption-investment problem, the linear-quadratic regulator, and a constrained consumption-investment problem with stochastic volatility. As benchmarks, we compare our results with those obtained using backward partial differential equation (PDE) methods and Physics-Informed Neural Networks (PINNs).
{"title":"Optimal control by policy iterations and constrained Gaussian process regressions","authors":"Donatien Hainaut , Jean-Loup Dupret","doi":"10.1016/j.ejor.2025.12.001","DOIUrl":"10.1016/j.ejor.2025.12.001","url":null,"abstract":"<div><div>This article introduces a novel iterative algorithm that combines policy iterations with Gaussian process regressions to solve stochastic control problems. At each iteration, an approximate value function is updated, followed by an improvement of the control strategy. State variables are sampled both within the interior domain and on the terminal boundary of the Hamilton-Jacobi-Bellman (HJB) equation. The approximate value function, interpreted as the solution to the HJB equation, is obtained by fitting a constrained regression model. This regression function aligns with the terminal utility on the boundary sample and satisfies the HJB equation on the interior sample. Assuming the regression function follows a Gaussian process, we derive closed-form approximations for both the value function and its derivatives. A numerical illustration demonstrates the efficiency of the proposed method in solving the consumption-investment problem, the linear-quadratic regulator, and a constrained consumption-investment problem with stochastic volatility. As benchmarks, we compare our results with those obtained using backward partial differential equation (PDE) methods and Physics-Informed Neural Networks (PINNs).</div></div>","PeriodicalId":55161,"journal":{"name":"European Journal of Operational Research","volume":"330 2","pages":"Pages 525-539"},"PeriodicalIF":6.0,"publicationDate":"2026-04-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145689306","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"管理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-04-16Epub Date: 2025-12-11DOI: 10.1016/j.ejor.2025.11.032
Sigrún Andradóttir, Hayriye Ayhan
We consider a Markovian queueing system with two types of customers (basic and advanced) and two types of servers (regular and specialist) in the presence of customer classification errors. We assume that there are always both types of customers waiting for service. When an advanced customer is misclassified as a basic customer, he needs to be served by a specialist after being served by a regular server. Our objective is to determine the dynamic assignment of the specialists between advanced and misclassified customers that maximizes the long-run average profit. We consider two versions of the problem that differ depending on whether the misclassified customers experience service continuity (the regular servers stay with misclassified customers while they wait for specialists, preventing the regular servers from serving other basic customers) or not (the regular servers continue serving other basic customers while misclassified customers wait for specialists). For both versions of the problem, we first characterize the optimal assignment of the specialists and then investigate how the optimal long-run average profit depends on the misclassification probability. We provide examples of systems where the optimal long-run average profit is not monotone in the misclassification probability, which is counter intuitive as one would expect misclassification to have a negative impact on system performance. We conclude our analysis by identifying under what conditions it is more profitable to serve customers with or without service continuity.
{"title":"Optimal server control with Two Customer Classes and Classification Errors","authors":"Sigrún Andradóttir, Hayriye Ayhan","doi":"10.1016/j.ejor.2025.11.032","DOIUrl":"10.1016/j.ejor.2025.11.032","url":null,"abstract":"<div><div>We consider a Markovian queueing system with two types of customers (basic and advanced) and two types of servers (regular and specialist) in the presence of customer classification errors. We assume that there are always both types of customers waiting for service. When an advanced customer is misclassified as a basic customer, he needs to be served by a specialist after being served by a regular server. Our objective is to determine the dynamic assignment of the specialists between advanced and misclassified customers that maximizes the long-run average profit. We consider two versions of the problem that differ depending on whether the misclassified customers experience service continuity (the regular servers stay with misclassified customers while they wait for specialists, preventing the regular servers from serving other basic customers) or not (the regular servers continue serving other basic customers while misclassified customers wait for specialists). For both versions of the problem, we first characterize the optimal assignment of the specialists and then investigate how the optimal long-run average profit depends on the misclassification probability. We provide examples of systems where the optimal long-run average profit is not monotone in the misclassification probability, which is counter intuitive as one would expect misclassification to have a negative impact on system performance. We conclude our analysis by identifying under what conditions it is more profitable to serve customers with or without service continuity.</div></div>","PeriodicalId":55161,"journal":{"name":"European Journal of Operational Research","volume":"330 2","pages":"Pages 512-524"},"PeriodicalIF":6.0,"publicationDate":"2026-04-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145731381","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"管理学","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