Omega-international Journal of Management Science最新文献_第4页

IF 7.2 2区管理学 Q1 MANAGEMENT

Omega-international Journal of Management Science

Pub Date : 2025-11-19 DOI: 10.1016/j.omega.2025.103475

Yishan Han , Biao Xu , Yao Wang , Shanxing Gao

Top-

K

recommendation involves inferring latent user preferences and generating personalized recommendations accordingly, which is now ubiquitous in various decision systems. Nonetheless, recommender systems usually suffer from severe popularity bias, leading to the over-recommendation of popular items. Such a bias deviates from the central aim of reflecting user preference faithfully, compromising both customer satisfaction and retailer profits. Despite the prevalence, existing methods tackling popularity bias still have limitations due to the considerable accuracy-debias tradeoff and the sensitivity to extensive parameter selection, further exacerbated by the extreme sparsity in positive user-item interactions.

In this paper, we present a Popularity-aware top-

K

recommendation algorithm integrating multi-behavior Side Information (PopSI), aiming to enhance recommendation accuracy and debias performance simultaneously. Specifically, by leveraging multiple user feedback that mirrors similar user preferences and formulating it as a three-dimensional tensor, PopSI can utilize all slices to capture the desiring user preferences effectively. Subsequently, we introduced a novel orthogonality constraint to refine the estimated item feature space, enforcing it to be invariant to item popularity features thereby addressing our model’s sensitivity to popularity bias. Comprehensive experiments on real-world e-commerce datasets demonstrate the general improvements of PopSI over state-of-the-art debias methods with a marginal accuracy-debias tradeoff and scalability to practical applications.

Top-K推荐涉及推断潜在的用户偏好并相应地生成个性化推荐，这在各种决策系统中普遍存在。尽管如此，推荐系统通常会遭受严重的流行偏见，导致过度推荐热门产品。这种偏差偏离了忠实反映用户偏好的中心目标，损害了顾客满意度和零售商利润。尽管流行，现有的解决流行偏差的方法仍然有局限性，由于相当大的准确性-偏差权衡和对广泛参数选择的敏感性，进一步加剧了积极用户-项目交互的极端稀疏性。本文提出了一种融合多行为侧信息（PopSI）的人气感知top-K推荐算法，旨在同时提高推荐精度和推荐性能。具体来说，通过利用反映类似用户偏好的多个用户反馈，并将其形成三维张量，PopSI可以利用所有切片来有效地捕获期望的用户偏好。随后，我们引入了一种新的正交性约束来改进估计的物品特征空间，使其对物品流行度特征不变，从而解决了我们的模型对流行度偏差的敏感性。在真实世界的电子商务数据集上进行的综合实验表明，PopSI比最先进的debias方法有了总体改进，并在边际精度-debias权衡和实际应用的可扩展性方面进行了改进。

{"title":"Towards popularity-aware recommendation: A multi-behavior enhanced framework with orthogonality constraint","authors":"Yishan Han , Biao Xu , Yao Wang , Shanxing Gao","doi":"10.1016/j.omega.2025.103475","DOIUrl":"10.1016/j.omega.2025.103475","url":null,"abstract":"<div><div>Top-<span><math><mi>K</mi></math></span> recommendation involves inferring latent user preferences and generating personalized recommendations accordingly, which is now ubiquitous in various decision systems. Nonetheless, recommender systems usually suffer from severe <em>popularity bias</em>, leading to the over-recommendation of popular items. Such a bias deviates from the central aim of reflecting user preference faithfully, compromising both customer satisfaction and retailer profits. Despite the prevalence, existing methods tackling popularity bias still have limitations due to the considerable accuracy-debias tradeoff and the sensitivity to extensive parameter selection, further exacerbated by the extreme sparsity in positive user-item interactions.</div><div>In this paper, we present a <strong>Pop</strong>ularity-aware top-<span><math><mi>K</mi></math></span> recommendation algorithm integrating multi-behavior <strong>S</strong>ide <strong>I</strong>nformation (PopSI), aiming to enhance recommendation accuracy and debias performance simultaneously. Specifically, by leveraging multiple user feedback that mirrors similar user preferences and formulating it as a three-dimensional tensor, PopSI can utilize all slices to capture the desiring user preferences effectively. Subsequently, we introduced a novel orthogonality constraint to refine the estimated item feature space, enforcing it to be invariant to item popularity features thereby addressing our model’s sensitivity to popularity bias. Comprehensive experiments on real-world e-commerce datasets demonstrate the general improvements of PopSI over state-of-the-art debias methods with a marginal accuracy-debias tradeoff and scalability to practical applications.</div></div>","PeriodicalId":19529,"journal":{"name":"Omega-international Journal of Management Science","volume":"140 ","pages":"Article 103475"},"PeriodicalIF":7.2,"publicationDate":"2025-11-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145571183","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}

引用次数: 0

Adaptation strategies-based supply chain viability optimization under bidirectional ripple effects 双向连锁效应下基于适应策略的供应链生存力优化

IF 7.2 2区管理学 Q1 MANAGEMENT

Omega-international Journal of Management Science

Pub Date : 2025-11-17 DOI: 10.1016/j.omega.2025.103467

Wei Pu , Xiangbin Yan , Shuang Ma

Long-term disruptions can trigger ripple effects across the supply chain, impacting both upstream and downstream stakeholders. To mitigate the consequences of such bidirectional ripple effects, including financial losses, consumer dissatisfaction, and declines in supply chain performance, an efficient optimization framework is required to enhance supply chain viability (SCV) through the integration of agility, resilience, and sustainability.

To strengthen the adaptive capacity and long-term viability of supply chain networks, we propose an adaptation strategies-based optimization framework. These adaptation strategies include utilizing third-party logistics (3PL) and implementing self-healing mechanisms. We then develop a two-stage multi-period stochastic programming model incorporating these strategies. This model explicitly captures bidirectional ripple effects by integrating the forward and backward propagation of dynamic disruptions. Additionally, we develop an advanced multi-objective particle swarm optimization (AMOPSO) solution method for solving the two-stage multi-period stochastic programming.

Using real-world data from China’s small appliances industry during COVID-19, we demonstrate the applicability of the proposed optimization model in enhancing SCV and mitigating bidirectional ripple effects, and the efficiency and robustness of the developed AMOPSO solution method. The optimal results reveal improvements of 3.22% in agility, 17.03% in resilience, and 23.02% in sustainability. Thus, the proposed optimization framework can improve supply chain viability under bidirectional ripple effects. The framework, along with the developed adaptation strategies-based stochastic optimization model and AMOPSO solution method, provides a novel approach to supply chain viability optimization and offers a practical decision-support method to original equipment manufacturers (OEMs) under bidirectional ripple effects.

长期的中断会引发整个供应链的连锁反应，影响上游和下游的利益相关者。为了减轻这种双向连锁反应的后果，包括财务损失、消费者不满和供应链绩效下降，需要一个有效的优化框架，通过整合敏捷性、弹性和可持续性来提高供应链的可行性（SCV）。为了增强供应链网络的适应能力和长期生存能力，我们提出了一个基于适应策略的优化框架。这些适应策略包括利用第三方物流（3PL）和实施自我修复机制。然后，我们开发了一个包含这些策略的两阶段多周期随机规划模型。该模型通过整合动态中断的正向和反向传播来明确捕获双向涟漪效应。此外，针对两阶段多周期随机规划问题，提出了一种先进的多目标粒子群优化（AMOPSO）方法。利用2019冠状病毒病期间中国小家电行业的实际数据，我们证明了所提出的优化模型在增强SCV和减轻双向连锁反应方面的适用性，以及所开发的AMOPSO解决方法的效率和鲁棒性。优化结果表明，敏捷性提高3.22%，弹性提高17.03%，可持续性提高23.02%。因此，本文提出的优化框架可以提高供应链在双向连锁效应下的生存能力。该框架与已建立的基于自适应策略的随机优化模型和AMOPSO求解方法一起，为供应链生存力优化提供了一种新的途径，为双向连锁效应下的原始设备制造商（oem）提供了实用的决策支持方法。

{"title":"Adaptation strategies-based supply chain viability optimization under bidirectional ripple effects","authors":"Wei Pu , Xiangbin Yan , Shuang Ma","doi":"10.1016/j.omega.2025.103467","DOIUrl":"10.1016/j.omega.2025.103467","url":null,"abstract":"<div><div>Long-term disruptions can trigger ripple effects across the supply chain, impacting both upstream and downstream stakeholders. To mitigate the consequences of such bidirectional ripple effects, including financial losses, consumer dissatisfaction, and declines in supply chain performance, an efficient optimization framework is required to enhance supply chain viability (SCV) through the integration of agility, resilience, and sustainability.</div><div>To strengthen the adaptive capacity and long-term viability of supply chain networks, we propose an adaptation strategies-based optimization framework. These adaptation strategies include utilizing third-party logistics (3PL) and implementing self-healing mechanisms. We then develop a two-stage multi-period stochastic programming model incorporating these strategies. This model explicitly captures bidirectional ripple effects by integrating the forward and backward propagation of dynamic disruptions. Additionally, we develop an advanced multi-objective particle swarm optimization (AMOPSO) solution method for solving the two-stage multi-period stochastic programming.</div><div>Using real-world data from China’s small appliances industry during COVID-19, we demonstrate the applicability of the proposed optimization model in enhancing SCV and mitigating bidirectional ripple effects, and the efficiency and robustness of the developed AMOPSO solution method. The optimal results reveal improvements of 3.22% in agility, 17.03% in resilience, and 23.02% in sustainability. Thus, the proposed optimization framework can improve supply chain viability under bidirectional ripple effects. The framework, along with the developed adaptation strategies-based stochastic optimization model and AMOPSO solution method, provides a novel approach to supply chain viability optimization and offers a practical decision-support method to original equipment manufacturers (OEMs) under bidirectional ripple effects.</div></div>","PeriodicalId":19529,"journal":{"name":"Omega-international Journal of Management Science","volume":"140 ","pages":"Article 103467"},"PeriodicalIF":7.2,"publicationDate":"2025-11-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145616413","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}

引用次数: 0

A communication on the paper “An alternative weight sensitivity analysis for PROMETHEE II rankings” 关于论文“PROMETHEE II排名的另一种权重敏感性分析”的交流

IF 7.2 2区管理学 Q1 MANAGEMENT

Omega-international Journal of Management Science

Pub Date : 2025-11-16 DOI: 10.1016/j.omega.2025.103465

Evangelos Triantaphyllou , Richard O’Shea , Yves De Smet , Nguyen Anh Vu Doan

This short communication examines the description of a mixed integer programming approach first in traduced by Doan and De Smet [Doan, N.A.V. and De Smet, Y., 2018. An alternative weight sensitivity analysis for PROMETHEE II rankings. Omega, 80, 166–174] for performing a sensitivity analysis on criteria weights under an additive aggregation step. This work has already attracted considerable interest in the scientific community. However, the original MILP model suffers from some descriptive issues. In the present short communication these issues are identified and then rectified. The corrected MILP model is then extended to make it more versatile than the original one. It is also shown how it can become an integral part of an intelligent approach to multiple criteria decision analysis (MCDA) and thus become a valuable tool for decision making.

本文探讨了Doan和De Smet首次提出的混合整数规划方法的描述[Doan， N.A.V.和De Smet， Y., 2018]。PROMETHEE II排名的另一种权重敏感性分析。Omega, 80,166 - 174]用于在加性聚合步骤下对标准权重进行敏感性分析。这项工作已经引起了科学界相当大的兴趣。然而，原始的MILP模型存在一些描述性问题。在本简短的通讯中，查明了这些问题，然后加以纠正。然后对修正后的MILP模型进行扩展，使其比原始模型更通用。它还显示了它如何成为多标准决策分析（MCDA）的智能方法的一个组成部分，从而成为决策制定的有价值的工具。

引用次数: 0

UNHRD’s humanitarian support in South Asia via Multistage Stochastic Programming 联合国开发计划署通过多阶段随机规划对南亚的人道主义支持

IF 7.2 2区管理学 Q1 MANAGEMENT

Omega-international Journal of Management Science

Pub Date : 2025-11-15 DOI: 10.1016/j.omega.2025.103464

Ruoyu Hu, Douglas Alem, Aakil Caunhye

One of the main tasks of the United Nations Humanitarian Response Depot (UNHRD) relies on allocating relief aid to save people who suffer from disasters. This task is particularly challenging in areas like South Asia, where relief aid efforts are confronted with complex transportation conditions, significant socioeconomic disparities, and the frequent occurrence of disasters, not to mention that financial resources are often scarce. In this paper, we develop a novel Multistage Stochastic Programming model to help UNHRD support critical decisions regarding site selection and relief aid allocation. Differently from the main literature, where these decisions are often made within a two-stage paradigm, our three-stage perspective takes into account in-kind donation campaigns that are triggered depending on the disaster impact and its effects, and is paramount to improving the effectiveness and fairness of the disaster relief operation. Our objective function maximizes the effectiveness of the disaster relief operation, defined as the extent to which it fulfills the needs of the population. Considering that different regions often exhibit distinct coping capacities, the effectiveness measure also factors in a vulnerability score to encourage relief aid allocation to the most in-need populations. The overall results show the importance of in-kind donation to achieve a more equitable relief aid allocation plan and the benefit of targeting more vulnerable regions under severely scarce resources.

联合国人道主义反应仓库的主要任务之一是分配救济援助，以拯救遭受灾害的人。这项任务在南亚等地区尤其具有挑战性，在这些地区，救援工作面临着复杂的运输条件、巨大的社会经济差距和频繁发生的灾害，更不用说财政资源往往稀缺。在本文中，我们开发了一个新的多阶段随机规划模型，以帮助联合国难民署支持有关选址和救济援助分配的关键决策。与主流文献的两阶段决策不同，我们的三阶段视角考虑了根据灾害影响和影响而引发的实物捐赠活动，这对提高救灾行动的有效性和公平性至关重要。我们的目标职能是最大限度地提高救灾行动的有效性，其定义是救灾行动在多大程度上满足了人民的需要。考虑到不同地区往往表现出不同的应对能力，有效性措施还考虑到脆弱性得分，以鼓励向最需要的人口分配救济援助。总体结果表明，实物捐赠对于实现更公平的救援援助分配计划的重要性，以及在资源严重稀缺的情况下，针对更脆弱的地区的益处。

{"title":"UNHRD’s humanitarian support in South Asia via Multistage Stochastic Programming","authors":"Ruoyu Hu, Douglas Alem, Aakil Caunhye","doi":"10.1016/j.omega.2025.103464","DOIUrl":"10.1016/j.omega.2025.103464","url":null,"abstract":"<div><div>One of the main tasks of the United Nations Humanitarian Response Depot (UNHRD) relies on allocating relief aid to save people who suffer from disasters. This task is particularly challenging in areas like South Asia, where relief aid efforts are confronted with complex transportation conditions, significant socioeconomic disparities, and the frequent occurrence of disasters, not to mention that financial resources are often scarce. In this paper, we develop a novel Multistage Stochastic Programming model to help UNHRD support critical decisions regarding site selection and relief aid allocation. Differently from the main literature, where these decisions are often made within a two-stage paradigm, our three-stage perspective takes into account in-kind donation campaigns that are triggered depending on the disaster impact and its effects, and is paramount to improving the effectiveness and fairness of the disaster relief operation. Our objective function maximizes the effectiveness of the disaster relief operation, defined as the extent to which it fulfills the needs of the population. Considering that different regions often exhibit distinct coping capacities, the effectiveness measure also factors in a vulnerability score to encourage relief aid allocation to the most in-need populations. The overall results show the importance of in-kind donation to achieve a more equitable relief aid allocation plan and the benefit of targeting more vulnerable regions under severely scarce resources.</div></div>","PeriodicalId":19529,"journal":{"name":"Omega-international Journal of Management Science","volume":"140 ","pages":"Article 103464"},"PeriodicalIF":7.2,"publicationDate":"2025-11-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145571184","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}

引用次数: 0

Multi-driver transportation scheduling for improving supply chain resilience 提高供应链弹性的多驾驶员运输调度

IF 7.2 2区管理学 Q1 MANAGEMENT

Omega-international Journal of Management Science

Pub Date : 2025-11-13 DOI: 10.1016/j.omega.2025.103461

Shaojun Lu , Yiyu Song , Min Kong , Chaoming Hu , Amir M. Fathollahi-Fard , Maxim A. Dulebenets

Optimizing transportation scheduling enhances the flexibility of resource allocation and transport operations, thereby reducing delays and costs while improving the resilience of supply chains. This study investigates a transportation scheduling problem aimed at minimizing the maximum completion time, incorporating key real-world considerations such as multiple drivers, loading and unloading times, round-trip transportation, distributed logistics centers, and the deterioration effect. Several structural properties of the problem are derived through a comprehensive preliminary analysis. Building on these properties, an exact algorithm for task sequencing is developed, and a Mixed-Integer Linear Programming model is formulated. A lower bound for the problem is also established. Given the NP-hard nature of the problem, we propose an enhanced Variable Neighborhood Search (VNS) algorithm, which integrates the exact algorithm with four neighborhood structures to accelerate convergence and improve solution quality. Experimental results indicate that the proposed intelligent algorithm significantly outperforms five state-of-the-art metaheuristics in both convergence speed and solution quality. This study further integrates deep learning to predict the best runtime of the proposed intelligent algorithm across problems of varying scales, which can reduce computational time in practical optimization scenarios. Sensitivity analysis highlights the critical influence of normal transportation times and the deterioration coefficient on scheduling performance, offering valuable theoretical insights for supply chain management. The findings of this research contribute to optimizing transportation task scheduling, enhancing supply chain resilience, and promoting sustainable development goals in supply chain management.

优化运输调度可以提高资源配置和运输作业的灵活性，从而减少延误和成本，同时提高供应链的弹性。本文研究了一个以最小化最大完工时间为目标的运输调度问题，并结合了多个驾驶员、装卸时间、往返运输、分布式物流中心和劣化效应等关键现实考虑因素。通过全面的初步分析，导出了问题的几个结构性质。在此基础上，提出了一种精确的任务排序算法，并建立了混合整数线性规划模型。并给出了问题的下界。针对该问题的NP-hard特性，本文提出了一种增强的可变邻域搜索（VNS）算法，该算法将精确算法与四种邻域结构相结合，以加速收敛并提高解的质量。实验结果表明，该算法在收敛速度和解质量上都明显优于五种最先进的元启发式算法。本研究进一步整合深度学习来预测所提出的智能算法在不同规模问题中的最佳运行时间，从而减少实际优化场景中的计算时间。敏感性分析强调了正常运输时间和恶化系数对调度绩效的重要影响，为供应链管理提供了有价值的理论见解。研究结果有助于优化运输任务调度，增强供应链弹性，促进供应链管理的可持续发展目标。

{"title":"Multi-driver transportation scheduling for improving supply chain resilience","authors":"Shaojun Lu , Yiyu Song , Min Kong , Chaoming Hu , Amir M. Fathollahi-Fard , Maxim A. Dulebenets","doi":"10.1016/j.omega.2025.103461","DOIUrl":"10.1016/j.omega.2025.103461","url":null,"abstract":"<div><div>Optimizing transportation scheduling enhances the flexibility of resource allocation and transport operations, thereby reducing delays and costs while improving the resilience of supply chains. This study investigates a transportation scheduling problem aimed at minimizing the maximum completion time, incorporating key real-world considerations such as multiple drivers, loading and unloading times, round-trip transportation, distributed logistics centers, and the deterioration effect. Several structural properties of the problem are derived through a comprehensive preliminary analysis. Building on these properties, an exact algorithm for task sequencing is developed, and a Mixed-Integer Linear Programming model is formulated. A lower bound for the problem is also established. Given the NP-hard nature of the problem, we propose an enhanced Variable Neighborhood Search (VNS) algorithm, which integrates the exact algorithm with four neighborhood structures to accelerate convergence and improve solution quality. Experimental results indicate that the proposed intelligent algorithm significantly outperforms five state-of-the-art metaheuristics in both convergence speed and solution quality. This study further integrates deep learning to predict the best runtime of the proposed intelligent algorithm across problems of varying scales, which can reduce computational time in practical optimization scenarios. Sensitivity analysis highlights the critical influence of normal transportation times and the deterioration coefficient on scheduling performance, offering valuable theoretical insights for supply chain management. The findings of this research contribute to optimizing transportation task scheduling, enhancing supply chain resilience, and promoting sustainable development goals in supply chain management.</div></div>","PeriodicalId":19529,"journal":{"name":"Omega-international Journal of Management Science","volume":"140 ","pages":"Article 103461"},"PeriodicalIF":7.2,"publicationDate":"2025-11-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145616410","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}

引用次数: 0

A Deep & Cross Network-based framework for online food delivery time prediction with driver-specific information 基于深度和交叉网络的基于司机特定信息的在线食品配送时间预测框架

IF 7.2 2区管理学 Q1 MANAGEMENT

Omega-international Journal of Management Science

Pub Date : 2025-11-12 DOI: 10.1016/j.omega.2025.103457

Genshen Fu , Yujie Chi , Li Zheng , Zuo-Jun Max Shen

In online food delivery operations, accurate delivery time prediction is crucial, as it underpins effective resource allocation and ensures customer satisfaction. While prediction models can be trained using historical delivery data, there is significant room for improving accuracy, especially in incorporating driver-specific information. However, this task faces several challenges: (1) limited data availability and privacy concerns, (2) driver heterogeneity and data dispersion, and (3) tacit driver knowledge and feature engineering complexity. To tackle these issues, we introduce a Deep & Cross Network-based (DCN-based) framework. This framework utilizes limited driver-specific information and dispersed data to automate feature engineering and enhance prediction accuracy, enabling a more personalized and precise prediction process. It leverages both low-order feature interactions, captured by the Cross Network, and high-order interactions from the Deep Neural Network (DNN), effectively balancing interpretability and predictive power. Extensive experiments using real-world data from Zomato demonstrate that our approach with driver-specific information significantly outperforms both traditional and state-of-the-art models, achieving superior results across all regression accuracy metrics. The best performance yields a root mean square error (RMSE) of 3.6660, representing a 35.93% improvement over models without driver-specific information. Furthermore, the framework’s automatic feature engineering provides deeper insights into the interactions between driver information and external factors, offering a valuable tool for improving online food delivery operations.

在网上外卖业务中，准确的配送时间预测是至关重要的，因为它是有效分配资源和确保客户满意度的基础。虽然预测模型可以使用历史交付数据进行训练，但准确性仍有很大的提高空间，特别是在纳入驾驶员特定信息方面。然而，这项任务面临着几个挑战：(1)有限的数据可用性和隐私问题；(2)驱动的异质性和数据分散；(3)隐性驱动知识和特征工程的复杂性。为了解决这些问题，我们引入了一个基于深度跨网络（Deep & Cross Network-based, DCN-based）的框架。该框架利用有限的驾驶员特定信息和分散的数据来自动化特征工程并提高预测精度，从而实现更加个性化和精确的预测过程。它利用了交叉网络捕获的低阶特征交互和深度神经网络（DNN）的高阶交互，有效地平衡了可解释性和预测能力。使用来自Zomato的真实世界数据进行的大量实验表明，我们针对驾驶员特定信息的方法显著优于传统和最先进的模型，在所有回归精度指标上都取得了卓越的结果。最佳性能的均方根误差（RMSE）为3.6660，与没有特定驱动程序信息的模型相比，提高了35.93%。此外，该框架的自动特征工程可以更深入地了解驾驶员信息与外部因素之间的相互作用，为改进在线送餐业务提供了有价值的工具。

{"title":"A Deep & Cross Network-based framework for online food delivery time prediction with driver-specific information","authors":"Genshen Fu , Yujie Chi , Li Zheng , Zuo-Jun Max Shen","doi":"10.1016/j.omega.2025.103457","DOIUrl":"10.1016/j.omega.2025.103457","url":null,"abstract":"<div><div>In online food delivery operations, accurate delivery time prediction is crucial, as it underpins effective resource allocation and ensures customer satisfaction. While prediction models can be trained using historical delivery data, there is significant room for improving accuracy, especially in incorporating driver-specific information. However, this task faces several challenges: (1) limited data availability and privacy concerns, (2) driver heterogeneity and data dispersion, and (3) tacit driver knowledge and feature engineering complexity. To tackle these issues, we introduce a Deep & Cross Network-based (DCN-based) framework. This framework utilizes limited driver-specific information and dispersed data to automate feature engineering and enhance prediction accuracy, enabling a more personalized and precise prediction process. It leverages both low-order feature interactions, captured by the Cross Network, and high-order interactions from the Deep Neural Network (DNN), effectively balancing interpretability and predictive power. Extensive experiments using real-world data from Zomato demonstrate that our approach with driver-specific information significantly outperforms both traditional and state-of-the-art models, achieving superior results across all regression accuracy metrics. The best performance yields a root mean square error (RMSE) of 3.6660, representing a 35.93% improvement over models without driver-specific information. Furthermore, the framework’s automatic feature engineering provides deeper insights into the interactions between driver information and external factors, offering a valuable tool for improving online food delivery operations.</div></div>","PeriodicalId":19529,"journal":{"name":"Omega-international Journal of Management Science","volume":"139 ","pages":"Article 103457"},"PeriodicalIF":7.2,"publicationDate":"2025-11-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145525915","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}

引用次数: 0

Managing inventory and financing decisions under ambiguity 在不明确的情况下管理库存和融资决策

IF 7.2 2区管理学 Q1 MANAGEMENT

Omega-international Journal of Management Science

Pub Date : 2025-11-12 DOI: 10.1016/j.omega.2025.103460

Cheng Qian , Zhaolin Li , Qi Fu

This study proposes a robust optimization framework to address the persistent challenges faced by micro and small enterprises (MSEs) in raising capital due to high levels of demand ambiguity. We examine a robust newsvendor model in which the firm has insufficient initial capital and needs to raise capital from an external fund provider. Without knowing the precise demand distribution, both the firm and the fund provider adopt a max–min decision rule based on the mean and variance of the demand. The firm aims to maximize his expected worst-case profit by determining the production quantity, while the fund provider offers equity or loan financing, seeking a fair market-determined return on the contributed capital. We derive the robust production quantity and financing agreements under both equity and loan financing. We show that equity financing attains the system-optimal outcome under distributional ambiguity, and propose a simple formula for the robust interest rate under loan financing. We further generalize our analysis to consider collateral and initial capital, and extend the base model to a robust principal–agent setting where the firm can exert an unobservable effort to influence demand. In the latter case, we show that equity financing outperforms loan financing across a wide range of parameter values, contrary to the existing literature without demand ambiguity. Our analysis offers guidance for practitioners and policymakers seeking effective strategies to promote growth while safeguarding fund providers in the MSE sector.

本研究提出了一个稳健的优化框架，以解决微型和小型企业（mse）在筹集资金方面面临的持续挑战，这是由于高度的需求模糊性。我们研究了一个强大的报摊模型，其中公司没有足够的初始资本，需要从外部资金提供者筹集资金。在不知道确切需求分布的情况下，企业和资金提供者都采用基于需求均值和方差的最大最小决策规则。企业的目标是通过确定生产数量来最大化其预期最坏情况下的利润，而资金提供者提供股权或贷款融资，寻求市场决定的公平的出资回报。推导出了股权融资和贷款融资下的稳健生产数量和融资协议。我们证明了在分配不明确的情况下，股权融资达到了系统最优的结果，并提出了贷款融资下稳健利率的简单公式。我们进一步将我们的分析推广到考虑抵押品和初始资本，并将基本模型扩展到稳健的委托代理设置，在该设置中，企业可以施加不可观察的努力来影响需求。在后一种情况下，我们表明股权融资在广泛的参数值范围内优于贷款融资，与没有需求歧义的现有文献相反。我们的分析为从业者和政策制定者提供了指导，以寻求有效的策略来促进增长，同时保护MSE部门的资金提供者。

{"title":"Managing inventory and financing decisions under ambiguity","authors":"Cheng Qian , Zhaolin Li , Qi Fu","doi":"10.1016/j.omega.2025.103460","DOIUrl":"10.1016/j.omega.2025.103460","url":null,"abstract":"<div><div>This study proposes a robust optimization framework to address the persistent challenges faced by micro and small enterprises (MSEs) in raising capital due to high levels of demand ambiguity. We examine a robust newsvendor model in which the firm has insufficient initial capital and needs to raise capital from an external fund provider. Without knowing the precise demand distribution, both the firm and the fund provider adopt a max–min decision rule based on the mean and variance of the demand. The firm aims to maximize his expected worst-case profit by determining the production quantity, while the fund provider offers equity or loan financing, seeking a fair market-determined return on the contributed capital. We derive the robust production quantity and financing agreements under both equity and loan financing. We show that equity financing attains the system-optimal outcome under distributional ambiguity, and propose a simple formula for the robust interest rate under loan financing. We further generalize our analysis to consider collateral and initial capital, and extend the base model to a robust principal–agent setting where the firm can exert an unobservable effort to influence demand. In the latter case, we show that equity financing outperforms loan financing across a wide range of parameter values, contrary to the existing literature without demand ambiguity. Our analysis offers guidance for practitioners and policymakers seeking effective strategies to promote growth while safeguarding fund providers in the MSE sector.</div></div>","PeriodicalId":19529,"journal":{"name":"Omega-international Journal of Management Science","volume":"140 ","pages":"Article 103460"},"PeriodicalIF":7.2,"publicationDate":"2025-11-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145616286","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}

引用次数: 0

Information disclosure and pricing decisions in competitive waste treatment systems: An agent-based approach 竞争性废物处理系统中的信息披露和定价决策：基于代理的方法

IF 7.2 2区管理学 Q1 MANAGEMENT

Omega-international Journal of Management Science

Pub Date : 2025-11-11 DOI: 10.1016/j.omega.2025.103463

Junfei Hu , Zhe Tian , Liang Cui , Peng Zhou

Private operators are increasingly involved in municipal solid waste management worldwide, resulting in competitive waste treatment systems. In such competitive systems, the gate fee as a crucial revenue stream for private operators needs to be set appropriately to capture a larger share of waste stream and maximize profit. Assessing the impact of information disclosure on gate fee pricing decisions provides valuable insights for policy analysis and decision-making. This study proposes an agent-based competitive waste treatment model to analyze gate fee pricing decisions under disclosed information. The proposed model outperforms traditional methods such as game theory by considering both cooperation and competition relationships among multiple agents. The experience-weighted attraction algorithm is utilized to solve the proposed model, enabling collaborative learning behavior to be considered in the decision-making process, thereby making it suitable for a disclosed environment. We apply the proposed approach to examine the Shenzhen waste treatment market in China. It has been found that without information disclosure, operators may misjudge allocation rules, causing landfills to withdraw from competition and significantly raise gate fees in retaliation. Besides, disclosing market information contributes to optimizing gate fee decisions, reducing government expenditure, and improving waste allocation. Disclosing allocation rules emerges as the most effective policy for Shenzhen waste treatment market. These findings are expected to provide government agencies with comprehensive insights into gate fee pricing decisions under conditions of information disclosure.

私营经营者越来越多地参与世界各地的城市固体废物管理，从而形成了具有竞争性的废物处理系统。在这种竞争体制中，门票费作为私营经营者的重要收入来源，需要适当设置，以获取更大的废物流份额，实现利润最大化。评估信息披露对门票定价决策的影响为政策分析和决策提供了有价值的见解。本文提出了一个基于主体的竞争性废物处理模型来分析信息披露条件下的闸费定价决策。该模型考虑了多个智能体之间的合作和竞争关系，优于博弈论等传统方法。利用经验加权吸引算法求解所提出的模型，使决策过程中能够考虑协同学习行为，从而使其适用于公开环境。我们运用该方法考察了中国深圳的垃圾处理市场。研究发现，在信息不公开的情况下，经营者可能会误判分配规则，导致垃圾填埋场退出竞争，并大幅提高门票费作为报复。此外，披露市场信息有助于优化门票收费决策，减少政府支出，改善垃圾分配。披露分配规则成为深圳垃圾处理市场最有效的政策。这些研究结果有望为政府机构在信息公开条件下的门票定价决策提供全面的见解。

{"title":"Information disclosure and pricing decisions in competitive waste treatment systems: An agent-based approach","authors":"Junfei Hu , Zhe Tian , Liang Cui , Peng Zhou","doi":"10.1016/j.omega.2025.103463","DOIUrl":"10.1016/j.omega.2025.103463","url":null,"abstract":"<div><div>Private operators are increasingly involved in municipal solid waste management worldwide, resulting in competitive waste treatment systems. In such competitive systems, the gate fee as a crucial revenue stream for private operators needs to be set appropriately to capture a larger share of waste stream and maximize profit. Assessing the impact of information disclosure on gate fee pricing decisions provides valuable insights for policy analysis and decision-making. This study proposes an agent-based competitive waste treatment model to analyze gate fee pricing decisions under disclosed information. The proposed model outperforms traditional methods such as game theory by considering both cooperation and competition relationships among multiple agents. The experience-weighted attraction algorithm is utilized to solve the proposed model, enabling collaborative learning behavior to be considered in the decision-making process, thereby making it suitable for a disclosed environment. We apply the proposed approach to examine the Shenzhen waste treatment market in China. It has been found that without information disclosure, operators may misjudge allocation rules, causing landfills to withdraw from competition and significantly raise gate fees in retaliation. Besides, disclosing market information contributes to optimizing gate fee decisions, reducing government expenditure, and improving waste allocation. Disclosing allocation rules emerges as the most effective policy for Shenzhen waste treatment market. These findings are expected to provide government agencies with comprehensive insights into gate fee pricing decisions under conditions of information disclosure.</div></div>","PeriodicalId":19529,"journal":{"name":"Omega-international Journal of Management Science","volume":"139 ","pages":"Article 103463"},"PeriodicalIF":7.2,"publicationDate":"2025-11-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145526323","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}

引用次数: 0

Two-stage distributionally robust optimization approach for drone-supported facility location and post-disaster relief distribution 无人机支持下设施选址与灾后救援分配的两阶段分布鲁棒优化方法

IF 7.2 2区管理学 Q1 MANAGEMENT

Omega-international Journal of Management Science

Pub Date : 2025-11-11 DOI: 10.1016/j.omega.2025.103462

Pan Gao , Min Li , Zhongming Wu , Zhenzhen Zhang

This paper explores the drone-supported application in a two-stage capacitated facility location problem, focusing on the strategic planning and operational phases of humanitarian relief. The first stage involves selecting facility locations and allocating drones, while the second stage involves transporting relief supplies post-disaster. We address the uncertainty inherent in post-disaster demand by employing a two-stage distributionally robust optimization (DRO) framework. To characterize various distributions of uncertainty, two types of ambiguity sets are introduced to characterize the unknown demand distribution: the Wasserstein and the event-wise mean absolute deviation ambiguity set. Furthermore, the DRO problem under the Wasserstein ambiguity set is decomposed and then solved using a column-and-constraint generation algorithm, due to the computational intractability of enumerating dual vertices in practical settings. In contrast, for the DRO problem tied to the mean absolute deviation ambiguity set, an event-wise affine decision rule is utilized to handle the recourse problem. This leads to reforming the problem into mixed-integer linear programming models, enabling its solution using standard optimization solvers. Numerical results demonstrate the effectiveness of both approaches, with the DRO models delivering more reliable solutions in out-of-sample tests compared to other state-of-the-art models. Specifically, our DRO models significantly reduce unmet post-disaster demand and ensure smaller total cost fluctuations on both in-sample and out-of-sample tests.

本文以人道主义救援的战略规划和操作阶段为重点，探讨了无人机支持在两阶段有能力设施选址问题中的应用。第一阶段包括选择设施地点和分配无人机，第二阶段包括灾后救援物资的运输。我们通过采用两阶段分布鲁棒优化（DRO）框架来解决灾后需求中固有的不确定性。为了描述各种不确定性分布，引入了两种类型的模糊集来描述未知需求分布：Wasserstein模糊集和事件平均绝对偏差模糊集。此外，由于在实际设置中枚举双顶点的计算困难，对Wasserstein模糊集下的DRO问题进行了分解，然后使用列约束生成算法求解。相反，对于与平均绝对偏差模糊集相关的DRO问题，使用事件仿射决策规则来处理追索权问题。这导致将问题转化为混合整数线性规划模型，使其能够使用标准优化求解器进行解决。数值结果证明了这两种方法的有效性，与其他最先进的模型相比，DRO模型在样本外测试中提供了更可靠的解决方案。具体来说，我们的DRO模型显著减少了灾后未满足的需求，并确保样本内和样本外测试的总成本波动较小。

{"title":"Two-stage distributionally robust optimization approach for drone-supported facility location and post-disaster relief distribution","authors":"Pan Gao , Min Li , Zhongming Wu , Zhenzhen Zhang","doi":"10.1016/j.omega.2025.103462","DOIUrl":"10.1016/j.omega.2025.103462","url":null,"abstract":"<div><div>This paper explores the drone-supported application in a two-stage capacitated facility location problem, focusing on the strategic planning and operational phases of humanitarian relief. The first stage involves selecting facility locations and allocating drones, while the second stage involves transporting relief supplies post-disaster. We address the uncertainty inherent in post-disaster demand by employing a two-stage distributionally robust optimization (DRO) framework. To characterize various distributions of uncertainty, two types of ambiguity sets are introduced to characterize the unknown demand distribution: the Wasserstein and the event-wise mean absolute deviation ambiguity set. Furthermore, the DRO problem under the Wasserstein ambiguity set is decomposed and then solved using a column-and-constraint generation algorithm, due to the computational intractability of enumerating dual vertices in practical settings. In contrast, for the DRO problem tied to the mean absolute deviation ambiguity set, an event-wise affine decision rule is utilized to handle the recourse problem. This leads to reforming the problem into mixed-integer linear programming models, enabling its solution using standard optimization solvers. Numerical results demonstrate the effectiveness of both approaches, with the DRO models delivering more reliable solutions in out-of-sample tests compared to other state-of-the-art models. Specifically, our DRO models significantly reduce unmet post-disaster demand and ensure smaller total cost fluctuations on both in-sample and out-of-sample tests.</div></div>","PeriodicalId":19529,"journal":{"name":"Omega-international Journal of Management Science","volume":"139 ","pages":"Article 103462"},"PeriodicalIF":7.2,"publicationDate":"2025-11-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145525914","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}

引用次数: 0

Target Setting for the Digital Economy in China: A DEA Bargaining Approach with General Production Network Structure 中国数字经济目标设定：基于一般生产网络结构的DEA议价方法

IF 7.2 2区管理学 Q1 MANAGEMENT

Omega-international Journal of Management Science

Pub Date : 2025-11-08 DOI: 10.1016/j.omega.2025.103459

Ming-Miin Yu , Sebastián Lozano , Kok Fong See

Over the past two decades, China's digital economy (DE) has grown significantly, and this is mainly due to investments in digital infrastructure, technological advancements, and government support. Despite this growth, regional disparities and sustainability challenges persist. This study aims to develop a more realistic and equitable target-setting framework for improving DE performance across provinces in China. We propose a generalized Network Data Envelopment Analysis (NDEA) model with a Nash bargaining mechanism, which allows cooperative optimization of inputs, intermediate products, and outputs among stages and subprocesses within a network production structure. The methodological innovation lies in its ability to capture interdependencies among subprocesses and distinguish between desirable, undesirable, and neutral intermediate products, thereby integrating green development considerations. Our empirical results show substantial improvement potential across provinces, with differentiated targets in areas such as energy consumption, software income, and e-commerce sales. The proposed model not only advances methodological development in the NDEA but also provides policymakers with a practical tool for promoting balanced regional development and sustainable digital transformation in China.

在过去的二十年里，中国的数字经济（DE）显著增长，这主要是由于对数字基础设施的投资、技术进步和政府的支持。尽管有这种增长，但区域差异和可持续性挑战依然存在。本研究旨在建立一个更加现实和公平的目标设定框架，以提高中国各省的DE绩效。我们提出了一个具有纳什议价机制的广义网络数据包络分析（NDEA）模型，该模型允许在网络生产结构中的阶段和子过程之间协作优化投入、中间产品和产出。方法上的创新在于它能够捕捉子过程之间的相互依赖关系，并区分理想的、不理想的和中性的中间产品，从而整合绿色发展的考虑。我们的实证结果显示，各省之间存在巨大的改善潜力，在能源消耗、软件收入和电子商务销售等领域的目标存在差异。该模型不仅推动了NDEA方法的发展，而且为政策制定者提供了促进中国区域平衡发展和可持续数字化转型的实用工具。

{"title":"Target Setting for the Digital Economy in China: A DEA Bargaining Approach with General Production Network Structure","authors":"Ming-Miin Yu , Sebastián Lozano , Kok Fong See","doi":"10.1016/j.omega.2025.103459","DOIUrl":"10.1016/j.omega.2025.103459","url":null,"abstract":"<div><div>Over the past two decades, China's digital economy (DE) has grown significantly, and this is mainly due to investments in digital infrastructure, technological advancements, and government support. Despite this growth, regional disparities and sustainability challenges persist. This study aims to develop a more realistic and equitable target-setting framework for improving DE performance across provinces in China. We propose a generalized Network Data Envelopment Analysis (NDEA) model with a Nash bargaining mechanism, which allows cooperative optimization of inputs, intermediate products, and outputs among stages and subprocesses within a network production structure. The methodological innovation lies in its ability to capture interdependencies among subprocesses and distinguish between desirable, undesirable, and neutral intermediate products, thereby integrating green development considerations. Our empirical results show substantial improvement potential across provinces, with differentiated targets in areas such as energy consumption, software income, and e-commerce sales. The proposed model not only advances methodological development in the NDEA but also provides policymakers with a practical tool for promoting balanced regional development and sustainable digital transformation in China.</div></div>","PeriodicalId":19529,"journal":{"name":"Omega-international Journal of Management Science","volume":"141 ","pages":"Article 103459"},"PeriodicalIF":7.2,"publicationDate":"2025-11-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145788002","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}

引用次数: 0