Computers & Industrial Engineering最新文献_第9页

Cross-silo human training in operational assembly: Integrating machine feedback for enhanced efficiency 操作装配中的跨仓人力培训：整合机器反馈以提高效率

IF 6.5 1区工程技术 Q1 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Computers & Industrial Engineering

Pub Date : 2026-03-01 Epub Date: 2025-12-18 DOI: 10.1016/j.cie.2025.111774

Kosuke Nakamura , Taro Ueyama , Masafumi Nishimura , Takayuki Nakano , Takahiro Aoki , Yoshitaka Yamamoto

In low-volume, multi-product manufacturing, workers must respond quickly and flexibly to changes in various work operations. However, there is currently a shortage of skilled workers, necessitating an effective method for training a wide range of workers with diverse characteristics. In this study, we first evaluated the feasibility of machine learning (ML) models for recognising complex assembly works. We next constructed the Feedback Integrated Expert Level Description System (FIELDS), which incorporates an ML model and functions for data collection, management, and user feedback. FIELDS can collect real-time work data from an action camera attached to trainees, analyse their assembly work from the data using the ML model, and provide feedback based on the analysis result. We evaluated the effect of the feedback using three metrics, the number of reduced missing processes, the distance from the regular processes, and the total work time. The feedback from the ML model was shown to enhance the trainees’ awareness of their proficiency and foster improvement. This result reveals a potential power of machine feedback for improving the efficiency of worker training. Consequently, this study contributes to offer a visible solution to enhance productivity and adaptability through cross-silo worker training in manufacturing environments.

在小批量、多产品制造中，工人必须快速灵活地响应各种工作操作的变化。然而，目前技术工人短缺，需要一种有效的方法来培训具有不同特点的广泛工人。在这项研究中，我们首先评估了机器学习（ML）模型识别复杂装配工作的可行性。接下来，我们构建了反馈集成专家级描述系统（FIELDS），该系统结合了ML模型和数据收集、管理和用户反馈的功能。FIELDS可以从附着在学员身上的动作相机收集实时工作数据，使用ML模型分析他们的组装工作数据，并根据分析结果提供反馈。我们使用三个度量来评估反馈的效果，减少缺失过程的数量，与常规过程的距离，以及总工作时间。从机器学习模型的反馈显示，提高学员的意识，他们的熟练程度和促进改进。这一结果揭示了机器反馈在提高工人培训效率方面的潜在力量。因此，本研究有助于提供一个可见的解决方案，以提高生产力和适应性，通过跨筒仓工人培训在制造环境。

{"title":"Cross-silo human training in operational assembly: Integrating machine feedback for enhanced efficiency","authors":"Kosuke Nakamura , Taro Ueyama , Masafumi Nishimura , Takayuki Nakano , Takahiro Aoki , Yoshitaka Yamamoto","doi":"10.1016/j.cie.2025.111774","DOIUrl":"10.1016/j.cie.2025.111774","url":null,"abstract":"<div><div>In low-volume, multi-product manufacturing, workers must respond quickly and flexibly to changes in various work operations. However, there is currently a shortage of skilled workers, necessitating an effective method for training a wide range of workers with diverse characteristics. In this study, we first evaluated the feasibility of machine learning (ML) models for recognising complex assembly works. We next constructed the <em>Feedback Integrated Expert Level Description System</em> (FIELDS), which incorporates an ML model and functions for data collection, management, and user feedback. FIELDS can collect real-time work data from an action camera attached to trainees, analyse their assembly work from the data using the ML model, and provide feedback based on the analysis result. We evaluated the effect of the feedback using three metrics, the number of reduced missing processes, the distance from the regular processes, and the total work time. The feedback from the ML model was shown to enhance the trainees’ awareness of their proficiency and foster improvement. This result reveals a potential power of machine feedback for improving the efficiency of worker training. Consequently, this study contributes to offer a visible solution to enhance productivity and adaptability through cross-silo worker training in manufacturing environments.</div></div>","PeriodicalId":55220,"journal":{"name":"Computers & Industrial Engineering","volume":"213 ","pages":"Article 111774"},"PeriodicalIF":6.5,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145841670","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A proposed local search strategy and neighborhoods for solving a new variant of e-commerce districting problem 提出了一种解决新型电子商务分区问题的局部搜索策略和邻域

IF 6.5 1区工程技术 Q1 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Computers & Industrial Engineering

Pub Date : 2026-03-01 Epub Date: 2026-01-20 DOI: 10.1016/j.cie.2026.111843

Van Son Nguyen, Quang Dung Pham, Quoc Trung Bui, Thuy Chau Tran

A districting problem with multiple-activity balancing is a key operational challenge in e-commerce logistics, where a large geographical area is divided into smaller zones allocated to drivers. The objective is to group such zones into a number of balanced operating districts and then assign them to drivers based on specific planning criteria. Optimizing these districts enables e-commerce companies to reduce operational costs while ensuring high delivery service quality. This paper introduces a novel variant of the p-Median districting problem, incorporating real-world factors inspired by one of the largest e-commerce companies in Southeast Asia. Our work considers significant workload differences between small zones and the driver’s familiarity. We formulate the considered problem as a mixed-integer linear programming model and propose an efficient local search framework to solve it. The novelty of our approach lies in designing new neighborhood structures to address workload imbalances in non-adjacent districts. To evaluate our proposed algorithm, numerical experiments are performed on both randomly generated and real-world instances. The proposed algorithm outperforms the location–allocation method developed by our partner company and some state-of-the-art methods in the literature, delivering high-quality solutions within a reasonable time. Thus, it is well-suited for real-world applications.

多活动平衡的分区问题是电子商务物流中一个关键的运营挑战，在电子商务物流中，一个大的地理区域被划分成更小的区域，分配给司机。目标是将这些区域划分为若干平衡的运营区域，然后根据具体的规划标准将其分配给司机。优化这些地区可以使电子商务公司在确保高交付服务质量的同时降低运营成本。本文介绍了p-中位数划分问题的一个新变体，结合了东南亚最大的电子商务公司之一所启发的现实世界因素。我们的研究考虑了小区域和驾驶员熟悉程度之间的显著工作量差异。我们将所考虑的问题表述为一个混合整数线性规划模型，并提出了一个有效的局部搜索框架来求解它。我们的方法的新颖之处在于设计新的社区结构，以解决非邻近地区的工作量不平衡问题。为了评估我们提出的算法，在随机生成和现实世界的实例上进行了数值实验。该算法优于合作伙伴公司开发的位置分配方法和文献中一些最先进的方法，在合理的时间内提供高质量的解决方案。因此，它非常适合实际应用程序。

{"title":"A proposed local search strategy and neighborhoods for solving a new variant of e-commerce districting problem","authors":"Van Son Nguyen, Quang Dung Pham, Quoc Trung Bui, Thuy Chau Tran","doi":"10.1016/j.cie.2026.111843","DOIUrl":"10.1016/j.cie.2026.111843","url":null,"abstract":"<div><div>A districting problem with multiple-activity balancing is a key operational challenge in e-commerce logistics, where a large geographical area is divided into smaller zones allocated to drivers. The objective is to group such zones into a number of balanced operating districts and then assign them to drivers based on specific planning criteria. Optimizing these districts enables e-commerce companies to reduce operational costs while ensuring high delivery service quality. This paper introduces a novel variant of the p-Median districting problem, incorporating real-world factors inspired by one of the largest e-commerce companies in Southeast Asia. Our work considers significant workload differences between small zones and the driver’s familiarity. We formulate the considered problem as a mixed-integer linear programming model and propose an efficient local search framework to solve it. The novelty of our approach lies in designing new neighborhood structures to address workload imbalances in non-adjacent districts. To evaluate our proposed algorithm, numerical experiments are performed on both randomly generated and real-world instances. The proposed algorithm outperforms the location–allocation method developed by our partner company and some state-of-the-art methods in the literature, delivering high-quality solutions within a reasonable time. Thus, it is well-suited for real-world applications.</div></div>","PeriodicalId":55220,"journal":{"name":"Computers & Industrial Engineering","volume":"213 ","pages":"Article 111843"},"PeriodicalIF":6.5,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146037806","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Competition and cooperation evaluation for multi-modal railway network: A multi-leader–follower approach 多式联运铁路网络的竞争与合作评价：多领导-跟随者方法

IF 6.5 1区工程技术 Q1 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Computers & Industrial Engineering

Pub Date : 2026-03-01 Epub Date: 2025-12-31 DOI: 10.1016/j.cie.2025.111759

Fangsheng Wang , Pengling Wang , Hanchuan Pan , Yuanchun Huang , Nikola Bešinović , Andrea D’Ariano

The multi-modal railway network, comprising high-speed rail (HSR), intercity rail (ICR), suburban rail (SUR), and urban rail transit (URT), has been gaining increasing attention due to its reliability and social benefits. These different rail transit modes often cooperate to provide seamless multi-modal travel services for long-distance passengers, while simultaneously facing competition due to overlapping passenger demand. However, most existing studies on ticket pricing and train scheduling do not fully account for the competition–cooperation relationships among these four rail transit modes. To address this gap, this study presents a multi-leader–follower game model that integrates ticket pricing and train scheduling (including line planning and timetabling) while considering the competitive and cooperative interactions within the multi-modal railway network. The model incorporates a simulation-based passenger assignment approach at the lower level and a decision-making framework at the upper level, aiming to approximate a Nash equilibrium solution among the various railway operators. The bus system is introduced only to provide a realistic competitive background, preventing a purely railway-monopolized setting and allowing us to better analyze the cooperative–competitive strategies among the four railway systems. An improved Nash Q-learning algorithm is developed to iteratively determine the approximated Nash equilibrium solution for the proposed multi-leader–follower game model. The effectiveness of the proposed method is demonstrated in a case study based on the multi-modal railway network in Jiangsu Province and Shanghai, China. Our results show that the proposed method can effectively optimize both ticket pricing and train scheduling in the multi-modal railway network under various competition–cooperation scenarios. A viable cooperation strategy involves encouraging passengers with short-distance trips to use urban transport modes (such as SUR and URT) while reserving more available seats on intercity transport modes (like HSR and ICR) for long-distance passengers. This strategy helps optimize the overall efficiency of the multi-modal transportation system.

由高速铁路（HSR）、城际铁路（ICR）、城郊铁路（SUR）和城市轨道交通（URT）组成的多式联运铁路网络因其可靠性和社会效益而越来越受到关注。这些不同的轨道交通方式往往相互合作，为长途旅客提供无缝的多式联运服务，同时由于旅客需求重叠而面临竞争。然而，现有的大多数关于票价和列车调度的研究都没有充分考虑到这四种轨道交通方式之间的竞争合作关系。为了解决这一差距，本研究提出了一个多领导-追随者博弈模型，该模型将车票定价和列车调度（包括线路规划和调度）集成在一起，同时考虑了多式联运铁路网络中的竞争和合作互动。该模型结合了基于仿真的下层乘客分配方法和上层决策框架，旨在近似各铁路运营商之间的纳什均衡解决方案。公交系统的引入只是为了提供一个现实的竞争背景，避免了纯粹的铁路垄断设置，使我们能够更好地分析四个铁路系统之间的合作竞争策略。提出了一种改进的纳什q -学习算法，用于迭代确定多领导-追随者博弈模型的近似纳什均衡解。以江苏省和上海市的多式联运铁路网络为例，验证了该方法的有效性。研究结果表明，该方法可以有效地优化多模式铁路网络中各种竞争-合作场景下的票价和列车调度。一个可行的合作策略是鼓励短途出行的乘客使用城市交通方式（如高铁和城际轨道交通），同时为长途乘客保留更多城际交通方式（如高铁和ICR）的可用座位。这一策略有助于优化多式联运系统的整体效率。

{"title":"Competition and cooperation evaluation for multi-modal railway network: A multi-leader–follower approach","authors":"Fangsheng Wang , Pengling Wang , Hanchuan Pan , Yuanchun Huang , Nikola Bešinović , Andrea D’Ariano","doi":"10.1016/j.cie.2025.111759","DOIUrl":"10.1016/j.cie.2025.111759","url":null,"abstract":"<div><div>The multi-modal railway network, comprising high-speed rail (HSR), intercity rail (ICR), suburban rail (SUR), and urban rail transit (URT), has been gaining increasing attention due to its reliability and social benefits. These different rail transit modes often cooperate to provide seamless multi-modal travel services for long-distance passengers, while simultaneously facing competition due to overlapping passenger demand. However, most existing studies on ticket pricing and train scheduling do not fully account for the competition–cooperation relationships among these four rail transit modes. To address this gap, this study presents a multi-leader–follower game model that integrates ticket pricing and train scheduling (including line planning and timetabling) while considering the competitive and cooperative interactions within the multi-modal railway network. The model incorporates a simulation-based passenger assignment approach at the lower level and a decision-making framework at the upper level, aiming to approximate a Nash equilibrium solution among the various railway operators. The bus system is introduced only to provide a realistic competitive background, preventing a purely railway-monopolized setting and allowing us to better analyze the cooperative–competitive strategies among the four railway systems. An improved Nash Q-learning algorithm is developed to iteratively determine the approximated Nash equilibrium solution for the proposed multi-leader–follower game model. The effectiveness of the proposed method is demonstrated in a case study based on the multi-modal railway network in Jiangsu Province and Shanghai, China. Our results show that the proposed method can effectively optimize both ticket pricing and train scheduling in the multi-modal railway network under various competition–cooperation scenarios. A viable cooperation strategy involves encouraging passengers with short-distance trips to use urban transport modes (such as SUR and URT) while reserving more available seats on intercity transport modes (like HSR and ICR) for long-distance passengers. This strategy helps optimize the overall efficiency of the multi-modal transportation system.</div></div>","PeriodicalId":55220,"journal":{"name":"Computers & Industrial Engineering","volume":"213 ","pages":"Article 111759"},"PeriodicalIF":6.5,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146037804","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Multi-objective home healthcare vehicle routing problem with drone-caregiver cooperation mode 基于无人机-护理员合作模式的多目标家庭医疗车辆路径问题

IF 6.5 1区工程技术 Q1 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Computers & Industrial Engineering

Pub Date : 2026-03-01 Epub Date: 2025-11-26 DOI: 10.1016/j.cie.2025.111714

Yichen Lu , Jun Yang , Chao Yang , Rui Zheng

The global population is experiencing accelerated aging, leading to a surge in demand for home healthcare services. This paper proposes a drone-assisted caregiver service mode for home healthcare, and investigates the multi-objective home healthcare vehicle routing problem with drone-caregiver cooperation mode (MHHVRPDC). In MHHVRPDC, drones are used to undertake basic medical delivery tasks, thereby reducing the workload of caregivers. This innovative approach empowers caregivers to concentrate on delivering more effective in-person patient care. We develop a multi-objective mixed-integer linear programming model for MHHVRPDC, aiming to minimize service costs and maximize task allocation fairness, while deciding on the drone-caregiver synchronized service route problem and the task assignment problem. To solve MHHVRPDC, we design a hybrid multi-objective evolutionary algorithm (HMOEA). The HMOEA employs a three-dimensional chromosome encoding scheme based on service modality differentiation (drone vs. caregiver) and caregivers’ skill levels, improves fitness evaluation rules, and designs specialized mutation operators and four categories of local search operators. A set of numerical experiments prove that HMOEAS is highly competitive. Further, taking the home healthcare services from Guanggu Campus of Tongji Hospital as a sample case, we demonstrate that the drone-caregiver cooperation mode outperforms the traditional caregiver-only service mode in reducing costs and improving task allocation fairness. Diversity scenarios such as different time windows, skill levels of the caregivers in the team, and the varying travel speed of caregiver are also analyzed to provide insights for decision-making.

全球人口正在加速老龄化，导致对家庭保健服务的需求激增。提出了一种面向家庭医疗的无人机辅助护理员服务模式，并研究了基于无人机-护理员合作模式（MHHVRPDC）的多目标家庭医疗车辆路径问题。在MHHVRPDC中，使用无人机承担基本的医疗交付任务，从而减少了护理人员的工作量。这种创新的方法使护理人员能够专注于提供更有效的面对面患者护理。以最小化服务成本和最大化任务分配公平性为目标，建立了MHHVRPDC多目标混合整数线性规划模型，同时确定了无人机-看护人同步服务路线问题和任务分配问题。为了解决MHHVRPDC问题，设计了一种混合多目标进化算法（HMOEA）。HMOEA采用基于服务模式区分（无人机vs.看护者）和看护者技能水平的三维染色体编码方案，改进适应度评价规则，设计专门的突变算子和四类局部搜索算子。一组数值实验证明了HMOEAS具有很强的竞争力。进一步以同济医院光谷校区家庭医疗服务为例，论证了无人机-护理员合作模式在降低成本和提高任务分配公平性方面优于传统的护理员服务模式。还分析了不同的时间窗口、团队中护理人员的技能水平、护理人员的不同旅行速度等多样性场景，为决策提供见解。

{"title":"Multi-objective home healthcare vehicle routing problem with drone-caregiver cooperation mode","authors":"Yichen Lu , Jun Yang , Chao Yang , Rui Zheng","doi":"10.1016/j.cie.2025.111714","DOIUrl":"10.1016/j.cie.2025.111714","url":null,"abstract":"<div><div>The global population is experiencing accelerated aging, leading to a surge in demand for home healthcare services. This paper proposes a drone-assisted caregiver service mode for home healthcare, and investigates the multi-objective home healthcare vehicle routing problem with drone-caregiver cooperation mode (MHHVRPDC). In MHHVRPDC, drones are used to undertake basic medical delivery tasks, thereby reducing the workload of caregivers. This innovative approach empowers caregivers to concentrate on delivering more effective in-person patient care. We develop a multi-objective mixed-integer linear programming model for MHHVRPDC, aiming to minimize service costs and maximize task allocation fairness, while deciding on the drone-caregiver synchronized service route problem and the task assignment problem. To solve MHHVRPDC, we design a hybrid multi-objective evolutionary algorithm (HMOEA). The HMOEA employs a three-dimensional chromosome encoding scheme based on service modality differentiation (drone vs. caregiver) and caregivers’ skill levels, improves fitness evaluation rules, and designs specialized mutation operators and four categories of local search operators. A set of numerical experiments prove that HMOEAS is highly competitive. Further, taking the home healthcare services from Guanggu Campus of Tongji Hospital as a sample case, we demonstrate that the drone-caregiver cooperation mode outperforms the traditional caregiver-only service mode in reducing costs and improving task allocation fairness. Diversity scenarios such as different time windows, skill levels of the caregivers in the team, and the varying travel speed of caregiver are also analyzed to provide insights for decision-making.</div></div>","PeriodicalId":55220,"journal":{"name":"Computers & Industrial Engineering","volume":"213 ","pages":"Article 111714"},"PeriodicalIF":6.5,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145841688","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Demand prediction for bike-sharing systems: A spatial and semantic modeling approach for enhanced accuracy and operational efficiency 自行车共享系统的需求预测：提高准确性和运行效率的空间和语义建模方法

IF 6.5 1区工程技术 Q1 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Computers & Industrial Engineering

Pub Date : 2026-03-01 Epub Date: 2025-12-20 DOI: 10.1016/j.cie.2025.111775

Juntao Wu , Jiahui Feng , Jie Fang , Hefu Liu

The exponential growth of Bike-Sharing Systems (BSS) has introduced complex challenges in supply–demand management, where imbalances frequently lead to resource wastage and reduced user satisfaction. While Graph Neural Networks (GNNs) have become a mainstream tool for demand forecasting, existing methodologies predominantly rely on static geographic proximity, failing to capture the latent semantic dependencies driven by actual riding behaviors. To bridge this gap, this paper proposes a novel Spatial-Semantic Graph Attention Neural Network (SSGAN). Unlike traditional models, SSGAN constructs a semantic adjacency matrix using DTW to quantify the shape similarity between station inflow and outflow patterns, thereby capturing non-Euclidean correlations beyond physical distance. Furthermore, a Gated Multi-Head Attention mechanism is designed to dynamically weigh these semantic relationships by integrating external covariates (e.g., weather), allowing the model to adapt to time-varying contexts. Crucially, to align prediction accuracy with decision effectiveness, the model employs a dual-stream architecture that fuses inflow and outflow features to better reflect net inventory changes. Empirical experiments on large-scale real-world datasets from Citi Bike and Divvy demonstrate that SSGAN not only achieves state-of-the-art prediction accuracy but also significantly reduces operational costs compared to baseline models. This study provides a generalized, decision-oriented computerized methodology for optimizing BSS rebalancing operations.

共享单车系统（BSS）的指数级增长给供需管理带来了复杂的挑战，其中不平衡经常导致资源浪费和用户满意度降低。虽然图神经网络（gnn）已经成为需求预测的主流工具，但现有的方法主要依赖于静态地理邻近性，无法捕获由实际骑行行为驱动的潜在语义依赖性。为了弥补这一缺陷，本文提出了一种新的空间语义图注意神经网络（SSGAN）。与传统模型不同，SSGAN使用DTW构建语义邻接矩阵来量化站点流入和流出模式之间的形状相似性，从而捕获超越物理距离的非欧几里得相关性。此外，设计了一个门控多头注意机制，通过整合外部协变量（如天气）来动态权衡这些语义关系，使模型能够适应时变的上下文。至关重要的是，为了使预测准确性与决策有效性保持一致，该模型采用了双流架构，融合了流入和流出特征，以更好地反映净库存变化。来自Citi Bike和Divvy的大规模真实数据集的实证实验表明，与基线模型相比，SSGAN不仅达到了最先进的预测精度，而且显著降低了运营成本。本研究为优化BSS再平衡操作提供了一种通用的、决策导向的计算机化方法。

{"title":"Demand prediction for bike-sharing systems: A spatial and semantic modeling approach for enhanced accuracy and operational efficiency","authors":"Juntao Wu , Jiahui Feng , Jie Fang , Hefu Liu","doi":"10.1016/j.cie.2025.111775","DOIUrl":"10.1016/j.cie.2025.111775","url":null,"abstract":"<div><div>The exponential growth of Bike-Sharing Systems (BSS) has introduced complex challenges in supply–demand management, where imbalances frequently lead to resource wastage and reduced user satisfaction. While Graph Neural Networks (GNNs) have become a mainstream tool for demand forecasting, existing methodologies predominantly rely on static geographic proximity, failing to capture the latent semantic dependencies driven by actual riding behaviors. To bridge this gap, this paper proposes a novel Spatial-Semantic Graph Attention Neural Network (SSGAN). Unlike traditional models, SSGAN constructs a semantic adjacency matrix using DTW to quantify the shape similarity between station inflow and outflow patterns, thereby capturing non-Euclidean correlations beyond physical distance. Furthermore, a Gated Multi-Head Attention mechanism is designed to dynamically weigh these semantic relationships by integrating external covariates (e.g., weather), allowing the model to adapt to time-varying contexts. Crucially, to align prediction accuracy with decision effectiveness, the model employs a dual-stream architecture that fuses inflow and outflow features to better reflect net inventory changes. Empirical experiments on large-scale real-world datasets from Citi Bike and Divvy demonstrate that SSGAN not only achieves state-of-the-art prediction accuracy but also significantly reduces operational costs compared to baseline models. This study provides a generalized, decision-oriented computerized methodology for optimizing BSS rebalancing operations.</div></div>","PeriodicalId":55220,"journal":{"name":"Computers & Industrial Engineering","volume":"213 ","pages":"Article 111775"},"PeriodicalIF":6.5,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145841617","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Beyond static schedules: Dynamic maintenance optimization with double deep reinforcement learning 超越静态时间表：动态维护优化与双重深度强化学习

IF 6.5 1区工程技术 Q1 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Computers & Industrial Engineering

Pub Date : 2026-03-01 Epub Date: 2026-01-10 DOI: 10.1016/j.cie.2026.111813

Allan Jonathan da Silva , Luís Domingues Tomé Jardim Tarrataca , Leonardo Fagundes de Mello , Fabricio Maione Tenório , Rodrigo Rodrigues de Freitas , Felipe do Carmo Amorim , Marcio Antelio Neves da Silva , Cintia Machado de Oliveira

This study presents an adaptive framework for dynamic preventive maintenance optimization based on the Double Deep Q-Network (DDQN) algorithm. The objective is to learn cost-optimal preventive maintenance policies under stochastic and partially observable failure behavior, relying solely on observed failure and maintenance events rather than condition-monitoring data or known degradation models. Equipment hazard function is modeled using non-homogeneous Poisson processes, including power-law and bathtub models, while maintenance actions follow variable restoration levels defined through the proportional age-reduction model. Training is performed on simulated failure trajectories using a standard workstation in under two hours, and the trained agent performs inference nearly instantaneously.

Results demonstrate that the DDQN-based adaptive policy consistently outperforms analytical periodic and static benchmarks, as well as a dynamic genetic algorithm and a standard reinforcement learning implementation, by achieving lower average maintenance costs and reduced variability across a wide range of corrective-to-preventive cost ratios. The method remains robust under perturbed and uncertain hazard conditions, maintaining stable performance without retraining.

These findings highlight the potential of the proposed DDQN approach as a computationally efficient and generalizable tool for reliability-centered maintenance optimization, capable of adapting to stochastic cost structures and cumulative corrective effects while operating effectively in data-limited industrial environments.

提出了一种基于双深度q网络（DDQN）算法的动态预防性维修优化自适应框架。目标是在随机和部分可观察到的故障行为下学习成本最优的预防性维护策略，仅依赖于观察到的故障和维护事件，而不是状态监测数据或已知的退化模型。设备危险函数使用非齐次泊松过程建模，包括幂律模型和浴缸模型，而维护行动遵循通过比例年龄减少模型定义的可变恢复水平。在两个小时内使用标准工作站在模拟故障轨迹上进行训练，训练后的智能体几乎可以立即执行推理。结果表明，基于ddqn的自适应策略始终优于分析周期性和静态基准，以及动态遗传算法和标准强化学习实现，通过实现更低的平均维护成本和在广泛的纠正-预防成本比范围内减少可变性。该方法在扰动和不确定的危险条件下保持鲁棒性，无需再训练即可保持稳定的性能。这些发现突出了DDQN方法作为一种计算效率高、可推广的以可靠性为中心的维护优化工具的潜力，能够适应随机成本结构和累积纠正效应，同时在数据有限的工业环境中有效运行。

{"title":"Beyond static schedules: Dynamic maintenance optimization with double deep reinforcement learning","authors":"Allan Jonathan da Silva , Luís Domingues Tomé Jardim Tarrataca , Leonardo Fagundes de Mello , Fabricio Maione Tenório , Rodrigo Rodrigues de Freitas , Felipe do Carmo Amorim , Marcio Antelio Neves da Silva , Cintia Machado de Oliveira","doi":"10.1016/j.cie.2026.111813","DOIUrl":"10.1016/j.cie.2026.111813","url":null,"abstract":"<div><div>This study presents an adaptive framework for dynamic preventive maintenance optimization based on the Double Deep Q-Network (DDQN) algorithm. The objective is to learn cost-optimal preventive maintenance policies under stochastic and partially observable failure behavior, relying solely on observed failure and maintenance events rather than condition-monitoring data or known degradation models. Equipment hazard function is modeled using non-homogeneous Poisson processes, including power-law and bathtub models, while maintenance actions follow variable restoration levels defined through the proportional age-reduction model. Training is performed on simulated failure trajectories using a standard workstation in under two hours, and the trained agent performs inference nearly instantaneously.</div><div>Results demonstrate that the DDQN-based adaptive policy consistently outperforms analytical periodic and static benchmarks, as well as a dynamic genetic algorithm and a standard reinforcement learning implementation, by achieving lower average maintenance costs and reduced variability across a wide range of corrective-to-preventive cost ratios. The method remains robust under perturbed and uncertain hazard conditions, maintaining stable performance without retraining.</div><div>These findings highlight the potential of the proposed DDQN approach as a computationally efficient and generalizable tool for reliability-centered maintenance optimization, capable of adapting to stochastic cost structures and cumulative corrective effects while operating effectively in data-limited industrial environments.</div></div>","PeriodicalId":55220,"journal":{"name":"Computers & Industrial Engineering","volume":"213 ","pages":"Article 111813"},"PeriodicalIF":6.5,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145977751","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A comprehensive review of time between events control charts: models, applications, and future directions 事件之间的时间控制图的全面审查：模型，应用程序和未来的方向

IF 6.5 1区工程技术 Q1 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Computers & Industrial Engineering

Pub Date : 2026-03-01 Epub Date: 2025-12-27 DOI: 10.1016/j.cie.2025.111789

Mohammadreza Mirzaei Novin , Amirhossein Amiri , Philippe Castagliola

The increasing demand for high-quality processes has driven growing interest in control charts designed for monitoring rare events. Among these, Time Between Events control charts have emerged as powerful tools, yet a comprehensive literature review dedicated to this field has been lacking until now. This gap has limited the ability of researchers and practitioners to identify key methodological advances, unresolved challenges, and emerging research directions. To address this need, the present study provides the first extensive review of TBE control charts, covering 113 studies published between 2000 and 2025. A multi-dimensional classification framework is introduced, which organizes the literature according to distributional assumptions, monitoring techniques, performance metrics, monitoring phases, event polarity (positive vs. negative events), and data structures (univariate, multivariate, and combined TBE with amplitude). The review further analyzes publication outlets and presents comprehensive reference tables to support quick identification of relevant methods. Finally, the paper highlights critical research gaps—including limited work on positive events, nonparametric multivariate monitoring, adaptive and hybrid methods, and machine learning integration—and proposes a forward-looking agenda to advance statistical process monitoring in increasingly complex environments.

对高质量过程的日益增长的需求推动了对为监视罕见事件而设计的控制图的日益增长的兴趣。其中，事件间隔时间控制图已经成为一种强大的工具，但到目前为止，还缺乏专门针对该领域的全面文献综述。这一差距限制了研究人员和实践者识别关键方法进展、未解决的挑战和新兴研究方向的能力。为了满足这一需求，本研究首次对TBE控制图进行了广泛的回顾，涵盖了2000年至2025年间发表的113项研究。介绍了一个多维分类框架，该框架根据分布假设、监测技术、性能指标、监测阶段、事件极性（积极与消极事件）和数据结构（单变量、多变量和结合TBE与振幅）组织文献。本文进一步分析了出版渠道，并提供了全面的参考表格，以支持快速识别相关方法。最后，本文强调了关键的研究差距——包括在积极事件、非参数多元监测、自适应和混合方法以及机器学习集成方面的有限工作——并提出了一个前瞻性议程，以促进在日益复杂的环境中进行统计过程监测。

{"title":"A comprehensive review of time between events control charts: models, applications, and future directions","authors":"Mohammadreza Mirzaei Novin , Amirhossein Amiri , Philippe Castagliola","doi":"10.1016/j.cie.2025.111789","DOIUrl":"10.1016/j.cie.2025.111789","url":null,"abstract":"<div><div>The increasing demand for high-quality processes has driven growing interest in control charts designed for monitoring rare events. Among these, Time Between Events control charts have emerged as powerful tools, yet a comprehensive literature review dedicated to this field has been lacking until now. This gap has limited the ability of researchers and practitioners to identify key methodological advances, unresolved challenges, and emerging research directions. To address this need, the present study provides the first extensive review of TBE control charts, covering 113 studies published between 2000 and 2025. A multi-dimensional classification framework is introduced, which organizes the literature according to distributional assumptions, monitoring techniques, performance metrics, monitoring phases, event polarity (positive vs. negative events), and data structures (univariate, multivariate, and combined TBE with amplitude). The review further analyzes publication outlets and presents comprehensive reference tables to support quick identification of relevant methods. Finally, the paper highlights critical research gaps—including limited work on positive events, nonparametric multivariate monitoring, adaptive and hybrid methods, and machine learning integration—and proposes a forward-looking agenda to advance statistical process monitoring in increasingly complex environments.</div></div>","PeriodicalId":55220,"journal":{"name":"Computers & Industrial Engineering","volume":"213 ","pages":"Article 111789"},"PeriodicalIF":6.5,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145977750","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A multi-type product design knowledge recommendation method for product conceptual design process 面向产品概念设计过程的多类型产品设计知识推荐方法

IF 6.5 1区工程技术 Q1 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Computers & Industrial Engineering

Pub Date : 2026-03-01 Epub Date: 2026-01-09 DOI: 10.1016/j.cie.2026.111812

Pengchao Wang, Jianjie Chu, Suihuai Yu, Bingkun Yuan, Xinyu Liu

The product conceptual design (PCD) process involves the integration, reasoning and reuse of multi-type product design knowledge (MTPDK) such as semantics, patents and case studies. To fully exploit the motivating potential of MTPDK, this paper proposes a recommendation method, which achieves a deeper integration between knowledge resources and the PCD process. First, a product design knowledge graph (PDKG) is constructed to represent semantic and patent knowledge through inter-entity relationships, while historical cases are encoded by connecting entities across layers via hyperedges. Next, the PCD process is formalized through the integration of Axiomatic Design (AD) and the Theory of Inventive Problem Solving (TRIZ), enabling a systematic analysis of knowledge requirements across different design stages. Based on the mapping of design problems across different dimensions, relevant MTPDK is recommended to designers. Specifically, a semantic activation diffusion algorithm is employed to support the zigzag mapping mechanism within AD, ensuring the rationality of the analysis and transformation processes. In parallel, patent knowledge novelty is evaluated to guide the application of TRIZ principles during the design matrix decoupling process. Furthermore, the case-matching degree is calculated to identify historical cases most relevant to the current design scenario, thereby facilitating adaptive design support. Subsequently, the proposed method is applied to the weeding equipment design process. The F₁ value of the knowledge recommendation result reaches 0.83, which verifies the feasibility and effectiveness of the proposed method. Finally, the comparative analyses demonstrate the superior performance of the proposed method.

产品概念设计（PCD）过程涉及对语义、专利和案例研究等多类型产品设计知识（MTPDK）的整合、推理和重用。为了充分挖掘MTPDK的激励潜力，本文提出了一种推荐方法，实现了知识资源与PCD过程的更深层次的整合。首先，构建产品设计知识图谱（PDKG），通过实体间关系表示语义知识和专利知识，同时通过超边跨层连接实体对历史案例进行编码。接下来，通过公理化设计（AD）和创造性问题解决理论（TRIZ）的整合，将PCD过程形式化，从而能够对不同设计阶段的知识需求进行系统分析。基于不同维度的设计问题映射，向设计人员推荐相关的MTPDK。具体而言，采用语义激活扩散算法支持AD内部的之字形映射机制，保证了分析和转换过程的合理性。同时，对专利知识的新颖性进行评估，以指导TRIZ原理在设计矩阵解耦过程中的应用。此外，计算案例匹配度以识别与当前设计场景最相关的历史案例，从而促进适应性设计支持。随后，将该方法应用于除草设备的设计过程中。知识推荐结果的F1值达到0.83，验证了所提方法的可行性和有效性。最后，通过对比分析验证了该方法的优越性。

{"title":"A multi-type product design knowledge recommendation method for product conceptual design process","authors":"Pengchao Wang, Jianjie Chu, Suihuai Yu, Bingkun Yuan, Xinyu Liu","doi":"10.1016/j.cie.2026.111812","DOIUrl":"10.1016/j.cie.2026.111812","url":null,"abstract":"<div><div>The product conceptual design (PCD) process involves the integration, reasoning and reuse of multi-type product design knowledge (MTPDK) such as semantics, patents and case studies. To fully exploit the motivating potential of MTPDK, this paper proposes a recommendation method, which achieves a deeper integration between knowledge resources and the PCD process. First, a product design knowledge graph (PDKG) is constructed to represent semantic and patent knowledge through inter-entity relationships, while historical cases are encoded by connecting entities across layers via hyperedges. Next, the PCD process is formalized through the integration of Axiomatic Design (AD) and the Theory of Inventive Problem Solving (TRIZ), enabling a systematic analysis of knowledge requirements across different design stages. Based on the mapping of design problems across different dimensions, relevant MTPDK is recommended to designers. Specifically, a semantic activation diffusion algorithm is employed to support the zigzag mapping mechanism within AD, ensuring the rationality of the analysis and transformation processes. In parallel, patent knowledge novelty is evaluated to guide the application of TRIZ principles during the design matrix decoupling process. Furthermore, the case-matching degree is calculated to identify historical cases most relevant to the current design scenario, thereby facilitating adaptive design support. Subsequently, the proposed method is applied to the weeding equipment design process. The <em>F</em><sub>1</sub> value of the knowledge recommendation result reaches 0.83, which verifies the feasibility and effectiveness of the proposed method. Finally, the comparative analyses demonstrate the superior performance of the proposed method.</div></div>","PeriodicalId":55220,"journal":{"name":"Computers & Industrial Engineering","volume":"213 ","pages":"Article 111812"},"PeriodicalIF":6.5,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145977749","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Efficient day-ahead energy scheduling in distribution systems via multi-objective symbiotic organism search 基于多目标共生生物搜索的配电系统日前能源调度

IF 6.5 1区工程技术 Q1 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Computers & Industrial Engineering

Pub Date : 2026-03-01 Epub Date: 2026-01-06 DOI: 10.1016/j.cie.2026.111804

Sibel Çevik Bektaş , Yeşim Aysel Baysal Aslanhan , İsmail Hakkı Altaş

An effective day-ahead planning strategy is pivotal for ensuring the economic, secure, and balanced operation of modern electricity grids. To address this challenge, various metaheuristic methods have been proposed for multi-objective day-ahead energy management, yet many suffer from scalability and convergence issues under realistic operating constraints. This study presents an efficient multi-objective optimization framework for day-ahead hourly optimal energy scheduling (DAHOES) in renewable-integrated distribution systems. The proposed framework employs the Fast Non-Dominated Sorting Multi-Objective Symbiotic Organism Search (FNSMOSOS) algorithm to minimize both active power losses and total operating costs. Following the optimization process, a fuzzy decision-making method is utilized to select a balanced solution from the generated Pareto front, ensuring that the final operation plan aligns with practical performance criteria. To reflect actual distribution system behavior, a modified five-bus distribution network comprising photovoltaic (PV) units, wind energy systems (WES), energy storage systems (ESS), and grid supply is modelled. In addition, realistic hourly demand profiles, renewable generation forecasts, and grid price signals are incorporated to ensure both theoretical optimality and practical feasibility. The proposed algorithm is compared with several other methods, and simulation results show that FNSMOSOS outperforms NSMOCS by 24.1% in HV and surpasses MOGWO, MOWOA, and MONNA by 56%, 117%, and 790%, respectively, demonstrating superior Pareto convergence and diversity. Overall, the results confirm that the proposed framework offers a scalable and effective decision-support tool for distribution system operators facing multi-criteria scheduling challenges in complex and uncertain power systems.

有效的日前规划策略是保证现代电网经济、安全、均衡运行的关键。为了解决这一挑战，人们提出了各种用于多目标日前能源管理的元启发式方法，但许多方法在实际操作约束下存在可扩展性和收敛性问题。提出了一种高效的多目标优化框架，用于可再生能源集成配电系统的日前小时最优能源调度。该框架采用快速非支配排序多目标共生生物搜索（FNSMOSOS）算法来最小化有功功率损耗和总运行成本。在优化过程之后，利用模糊决策方法从生成的Pareto front中选择一个平衡解，以确保最终的运行计划符合实际性能标准。为了反映实际的配电系统行为，对一个由光伏（PV）单元、风能系统（WES）、储能系统（ESS）和电网组成的改进的五总线配电网络进行了建模。此外，实际的小时需求概况，可再生能源发电预测和电网价格信号被纳入，以确保理论最优和实际可行性。仿真结果表明，FNSMOSOS算法的HV值比NSMOCS算法高24.1%，分别比MOGWO、MOWOA和MONNA算法高56%、117%和790%，具有较好的Pareto收敛性和多样性。总体而言，研究结果证实，该框架为配电系统运营商在复杂和不确定的电力系统中面临多准则调度挑战提供了一种可扩展和有效的决策支持工具。

{"title":"Efficient day-ahead energy scheduling in distribution systems via multi-objective symbiotic organism search","authors":"Sibel Çevik Bektaş , Yeşim Aysel Baysal Aslanhan , İsmail Hakkı Altaş","doi":"10.1016/j.cie.2026.111804","DOIUrl":"10.1016/j.cie.2026.111804","url":null,"abstract":"<div><div>An effective day-ahead planning strategy is pivotal for ensuring the economic, secure, and balanced operation of modern electricity grids. To address this challenge, various metaheuristic methods have been proposed for multi-objective day-ahead energy management, yet many suffer from scalability and convergence issues under realistic operating constraints. This study presents an efficient multi-objective optimization framework for day-ahead hourly optimal energy scheduling (DAHOES) in renewable-integrated distribution systems. The proposed framework employs the Fast Non-Dominated Sorting Multi-Objective Symbiotic Organism Search (FNSMOSOS) algorithm to minimize both active power losses and total operating costs. Following the optimization process, a fuzzy decision-making method is utilized to select a balanced solution from the generated Pareto front, ensuring that the final operation plan aligns with practical performance criteria. To reflect actual distribution system behavior, a modified five-bus distribution network comprising photovoltaic (PV) units, wind energy systems (WES), energy storage systems (ESS), and grid supply is modelled. In addition, realistic hourly demand profiles, renewable generation forecasts, and grid price signals are incorporated to ensure both theoretical optimality and practical feasibility. The proposed algorithm is compared with several other methods, and simulation results show that FNSMOSOS outperforms NSMOCS by 24.1% in HV and surpasses MOGWO, MOWOA, and MONNA by 56%, 117%, and 790%, respectively, demonstrating superior Pareto convergence and diversity. Overall, the results confirm that the proposed framework offers a scalable and effective decision-support tool for distribution system operators facing multi-criteria scheduling challenges in complex and uncertain power systems.</div></div>","PeriodicalId":55220,"journal":{"name":"Computers & Industrial Engineering","volume":"213 ","pages":"Article 111804"},"PeriodicalIF":6.5,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145927271","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Optimization of ambulance services sequencing and scheduling daily decisions with minimizing delay 优化救护车服务排序和调度日常决策，尽量减少延误

IF 6.5 1区工程技术 Q1 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Computers & Industrial Engineering

Pub Date : 2026-03-01 Epub Date: 2026-01-02 DOI: 10.1016/j.cie.2026.111803

Mohammad Javad Eslami , Mohsen Varmazyar

Ambulances, one of the essential resources in the emergency medical service (EMS), are crucial in transporting patients to hospitals and saving lives. This research addresses the ambulance service scheduling problem (ASSP) for daily planning decisions by minimizing total weighted tardiness. A mixed integer linear mathematical model for the research problem is developed. Since the research problem is shown to be NP-hard, two population-based genetic algorithm (GA) and particle swarm optimization (PSO), and two solution-based, simulated annealing (SA) and tabu search (TS) meta-heuristics are proposed to solve this problem. In addition, the Lagrangian relaxation (LR) and Benders decomposition methods are employed to find effective lower bounds. Random test problems with small, medium, and large sizes are generated and solved by the proposed algorithms to evaluate their performance. Numerical results show that the LR and Benders decomposition can find efficient lower bounds with approximately 4 % and 6 % gap rates, respectively. Furthermore, ANOVA and Tukey’s HSD tests indicate that the GA, PSO, and SA algorithms perform better in small-, medium-, and large-size problems, respectively. It is noticeable that the best-obtained meta-heuristic solutions have a gap rate of approximately 6.21 %, with the best-obtained lower bounds. Moreover, due to the ASSP problem’s stochastic nature, we develop a two-stage stochastic programming model by considering each mission’s weight and time under uncertainty. Additionally, considering enough scenarios, which in our research is 40, the optimal value can be closely approximated. The outputs of this research are employed for a real-world case study as well. Finally, some managerial and practical insights are discussed based on the results.

救护车是紧急医疗服务的重要资源之一，在将病人送往医院和挽救生命方面起着至关重要的作用。本研究通过最小化总加权延迟来解决救护车服务调度问题（ASSP）的日常规划决策。建立了研究问题的混合整数线性数学模型。针对研究问题具有np困难的特点，提出了两种基于种群的遗传算法（GA）和粒子群算法（PSO），以及两种基于解的模拟退火算法（SA）和禁忌搜索（TS）元启发式算法来解决该问题。此外，采用拉格朗日松弛（LR）和Benders分解方法寻找有效下界。通过生成和解决小、中、大尺寸的随机测试问题来评估算法的性能。数值结果表明，LR分解和Benders分解分别可以找到间隙率约为4%和6%的有效下界。此外，方差分析和Tukey的HSD检验表明，GA、PSO和SA算法分别在小型、中型和大型问题中表现更好。值得注意的是，最佳启发式解的间隙率约为6.21%，并具有最佳下界。此外，由于ASSP问题的随机性，我们建立了一个考虑不确定条件下每个任务的权值和时间的两阶段随机规划模型。此外，考虑到足够多的场景，在我们的研究中是40，最优值可以很接近。本研究的产出也用于现实世界的案例研究。最后，根据研究结果讨论了一些管理和实践见解。

{"title":"Optimization of ambulance services sequencing and scheduling daily decisions with minimizing delay","authors":"Mohammad Javad Eslami , Mohsen Varmazyar","doi":"10.1016/j.cie.2026.111803","DOIUrl":"10.1016/j.cie.2026.111803","url":null,"abstract":"<div><div>Ambulances, one of the essential resources in the emergency medical service (EMS), are crucial in transporting patients to hospitals and saving lives. This research addresses the ambulance service scheduling problem (ASSP) for daily planning decisions by minimizing total weighted tardiness. A mixed integer linear mathematical model for the research problem is developed. Since the research problem is shown to be NP-hard, two population-based genetic algorithm (GA) and particle swarm optimization (PSO), and two solution-based, simulated annealing (SA) and tabu search (TS) meta-heuristics are proposed to solve this problem. In addition, the Lagrangian relaxation (LR) and Benders decomposition methods are employed to find effective lower bounds. Random test problems with small, medium, and large sizes are generated and solved by the proposed algorithms to evaluate their performance. Numerical results show that the LR and Benders decomposition can find efficient lower bounds with approximately 4 % and 6 % gap rates, respectively. Furthermore, ANOVA and Tukey’s HSD tests indicate that the GA, PSO, and SA algorithms perform better in small-, medium-, and large-size problems, respectively. It is noticeable that the best-obtained meta-heuristic solutions have a gap rate of approximately 6.21 %, with the best-obtained lower bounds. Moreover, due to the ASSP problem’s stochastic nature, we develop a two-stage stochastic programming model by considering each mission’s weight and time under uncertainty. Additionally, considering enough scenarios, which in our research is 40, the optimal value can be closely approximated. The outputs of this research are employed for a real-world case study as well. Finally, some managerial and practical insights are discussed based on the results.</div></div>","PeriodicalId":55220,"journal":{"name":"Computers & Industrial Engineering","volume":"213 ","pages":"Article 111803"},"PeriodicalIF":6.5,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145927780","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0