Advanced Engineering Informatics最新文献_第3页

CRFPI-Net: context-aware risk feature perception and inference network for pixel-level urban traffic risk mapping CRFPI-Net：面向像素级城市交通风险映射的情境感知风险特征感知与推理网络

IF 9.9 1区工程技术 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Advanced Engineering Informatics

Pub Date : 2026-01-10 DOI: 10.1016/j.aei.2025.104299

Wentong Guo , Wenzhu Xu , Chengcheng Yang , Zhijian Zhao , Xi Gao , WenBin Yao , Sheng Jin

Urban traffic accidents result in significant casualties and property losses. Conducting traffic risk mapping and inference for urban areas provides substantial benefits for accident prevention as well as future planning and governance. However, pixel-level fine-grained inference of urban traffic risk maps remains challenging, primarily due to the complex layout of urban road networks, the temporal variability of traffic dynamics, and the heterogeneity of spatial semantic information. In this study, we propose an end-to-end Context-Aware Risk Feature Perception and Inference Network (CRFPI-Net) based on multimodal data to achieve fine-grained inference of urban traffic risk maps. In CRFPI-Net, three separate branches are designed to capture risk features from satellite remote sensing imagery, spatiotemporal traffic sequences, and area-of-interest (AOI) semantic information. The risk-aware features from each branch are integrated using a gated fusion mechanism to eliminate redundant information, and the fused features are further processed by context-aware multi-scale correlation analysis to reduce the adverse impact of heterogeneous variations in risk regions on risk perception. Finally, CRFPI-Net produces pixel-level inference maps of urban traffic accident risk, enabling effective and low-cost guidance for traffic accident prevention. The proposed model is quantitatively evaluated on real-world datasets and achieves state-of-the-art performance. Ablation experiments further demonstrate the rationality and effectiveness of the designed modules. The code and pretrained models for urban traffic risk mapping are publicly available at https://github.com/gwt-ZJU/CRFPI-Net.

城市交通事故造成重大人员伤亡和财产损失。对城市地区进行交通风险测绘和推断，为事故预防以及未来规划和治理提供了实质性的好处。然而，由于城市道路网络的复杂布局、交通动态的时间变异性和空间语义信息的异质性，城市交通风险地图的像素级细粒度推理仍然具有挑战性。在本研究中，我们提出了一个基于多模态数据的端到端上下文感知风险特征感知与推理网络（CRFPI-Net），以实现城市交通风险地图的细粒度推理。在CRFPI-Net中，设计了三个独立的分支来捕获来自卫星遥感图像、时空交通序列和兴趣区域（AOI）语义信息的风险特征。采用门控融合机制对各分支的风险感知特征进行融合，消除冗余信息，并通过上下文感知多尺度相关分析对融合特征进行进一步处理，降低风险区域异质变化对风险感知的不利影响。最后，CRFPI-Net生成城市交通事故风险的像素级推理地图，为交通事故预防提供有效和低成本的指导。所提出的模型在真实世界的数据集上进行了定量评估，并达到了最先进的性能。烧蚀实验进一步验证了所设计模块的合理性和有效性。城市交通风险地图的代码和预训练模型可在https://github.com/gwt-ZJU/CRFPI-Net上公开获取。

{"title":"CRFPI-Net: context-aware risk feature perception and inference network for pixel-level urban traffic risk mapping","authors":"Wentong Guo , Wenzhu Xu , Chengcheng Yang , Zhijian Zhao , Xi Gao , WenBin Yao , Sheng Jin","doi":"10.1016/j.aei.2025.104299","DOIUrl":"10.1016/j.aei.2025.104299","url":null,"abstract":"<div><div>Urban traffic accidents result in significant casualties and property losses. Conducting traffic risk mapping and inference for urban areas provides substantial benefits for accident prevention as well as future planning and governance. However, pixel-level fine-grained inference of urban traffic risk maps remains challenging, primarily due to the complex layout of urban road networks, the temporal variability of traffic dynamics, and the heterogeneity of spatial semantic information. In this study, we propose an end-to-end Context-Aware Risk Feature Perception and Inference Network (CRFPI-Net) based on multimodal data to achieve fine-grained inference of urban traffic risk maps. In CRFPI-Net, three separate branches are designed to capture risk features from satellite remote sensing imagery, spatiotemporal traffic sequences, and area-of-interest (AOI) semantic information. The risk-aware features from each branch are integrated using a gated fusion mechanism to eliminate redundant information, and the fused features are further processed by context-aware multi-scale correlation analysis to reduce the adverse impact of heterogeneous variations in risk regions on risk perception. Finally, CRFPI-Net produces pixel-level inference maps of urban traffic accident risk, enabling effective and low-cost guidance for traffic accident prevention. The proposed model is quantitatively evaluated on real-world datasets and achieves state-of-the-art performance. Ablation experiments further demonstrate the rationality and effectiveness of the designed modules. The code and pretrained models for urban traffic risk mapping are publicly available at <span><span>https://github.com/gwt-ZJU/CRFPI-Net</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":50941,"journal":{"name":"Advanced Engineering Informatics","volume":"71 ","pages":"Article 104299"},"PeriodicalIF":9.9,"publicationDate":"2026-01-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145978253","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

Enhancing industrial fault diagnosis via A Meta-learning: Zero-shot identification with constraint conditional variational autoencoder 基于元学习的工业故障诊断：约束条件变分自编码器零采样识别

IF 9.9 1区工程技术 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Advanced Engineering Informatics

Pub Date : 2026-01-10 DOI: 10.1016/j.aei.2025.104305

Li Zhu , Ching-Lien Liu , Junghui Chen

To address the challenges of zero-shot fault diagnosis in industrial systems, this study proposes a novel approach—Meta-CCVAE, a constrained conditional variational autoencoder (CCVAE) embedded within a meta-learning framework. By leveraging predefined attributes, Meta-CCVAE expands the representation space for unseen faults across three training phases, improving the quality of generated samples and minimizing overlap with known faults. To mitigate the limitations of scarce fault data, the framework incorporates both common and individual models: the common model enables knowledge sharing across fault classes, while the individual model ensures accurate fault characterization. This dual-model strategy reduces the impact of limited data and lowers computational costs. Experimental results on both simulated and real-world datasets validate the effectiveness of Meta-CCVAE, highlighting its potential for reliable zero-shot fault identification in industrial applications.

为了解决工业系统零故障诊断的挑战，本研究提出了一种新的方法-元CCVAE，一种嵌入元学习框架中的约束条件变分自编码器（CCVAE）。通过利用预定义的属性，Meta-CCVAE在三个训练阶段扩展了未见故障的表示空间，提高了生成样本的质量，并最大限度地减少了与已知故障的重叠。为了减轻故障数据稀缺的局限性，该框架结合了公共模型和单个模型：公共模型实现了故障类之间的知识共享，而单个模型确保了准确的故障表征。这种双模型策略减少了有限数据的影响并降低了计算成本。在模拟和真实数据集上的实验结果验证了Meta-CCVAE的有效性，突出了其在工业应用中可靠的零射击故障识别的潜力。

引用次数: 0

Tunable plasmonic absorption in metal–dielectric multilayers via FDTD simulations and an explainable machine learning approach 通过FDTD模拟和可解释的机器学习方法在金属介质多层中的可调谐等离子体吸收

IF 9.9 1区工程技术 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Advanced Engineering Informatics

Pub Date : 2026-01-09 DOI: 10.1016/j.aei.2026.104311

Emmanuel A. Bamidele

Plasmonic devices, fundamental to modern nanophotonics, exploit resonant interactions between light and free electrons in metals to achieve enhanced light trapping and electromagnetic field confinement. However, modeling their complex, nonlinear optical responses remain computationally intensive. In this work, we combine finite-difference time-domain (FDTD) simulations with machine learning (ML) to simulate and predict absorbed power behavior in multilayer plasmonic stacks composed of SiO₂, gold (Au), silver (Ag), and indium tin oxide (ITO). By varying Au and Ag thicknesses (10–50 nm) across a spectral range of 300–1500 nm, we generate spatial absorption maps and integrated power metrics from full-wave solutions to Maxwell’s equations. A multilayer perceptron models global absorption behavior with a mean absolute error (MAE) of 0.0953, while a convolutional neural network predicts spatial absorption distributions with an MAE of 0.0101. SHapley Additive exPlanations identify plasmonic layer thickness and excitation wavelength as dominant contributors to absorption, which peaks between 450 and 850 nm. Gold demonstrates broader and more sustained absorption compared to silver, although both metals exhibit reduced efficiency outside the resonance window. The integrated FDTD–ML framework accelerates plasmonic design while maintaining physical interpretability and predictive accuracy, enabling efficient exploration of tunable optical responses in multilayer nanophotonic systems for applications in optical sensing, photovoltaics, and device engineering.

等离子体器件是现代纳米光子学的基础，利用金属中光和自由电子之间的共振相互作用来实现增强的光捕获和电磁场约束。然而，模拟它们复杂的非线性光学响应仍然需要大量的计算。在这项工作中，我们将时域有限差分（FDTD）模拟与机器学习（ML）相结合，以模拟和预测由SiO2、金（Au）、银（Ag）和氧化铟锡（ITO）组成的多层等离子体堆叠的吸收功率行为。通过在300-1500 nm的光谱范围内改变Au和Ag的厚度（10-50 nm），我们从麦克斯韦方程的全波解中生成空间吸收图和集成功率指标。多层感知器模型全局吸收行为的平均绝对误差（MAE）为0.0953，而卷积神经网络预测空间吸收分布的平均绝对误差为0.0101。SHapley加性解释确定等离子体层厚度和激发波长是吸收的主要贡献者，其峰值在450和850 nm之间。与银相比，金表现出更广泛和更持久的吸收，尽管这两种金属在共振窗口外的效率都有所降低。集成的FDTD-ML框架加速了等离子体设计，同时保持了物理可解释性和预测准确性，能够有效地探索用于光学传感、光伏和器件工程应用的多层纳米光子系统中的可调谐光学响应。

{"title":"Tunable plasmonic absorption in metal–dielectric multilayers via FDTD simulations and an explainable machine learning approach","authors":"Emmanuel A. Bamidele","doi":"10.1016/j.aei.2026.104311","DOIUrl":"10.1016/j.aei.2026.104311","url":null,"abstract":"<div><div>Plasmonic devices, fundamental to modern nanophotonics, exploit resonant interactions between light and free electrons in metals to achieve enhanced light trapping and electromagnetic field confinement. However, modeling their complex, nonlinear optical responses remain computationally intensive. In this work, we combine finite-difference time-domain (FDTD) simulations with machine learning (ML) to simulate and predict absorbed power behavior in multilayer plasmonic stacks composed of SiO<sub>2</sub>, gold (Au), silver (Ag), and indium tin oxide (ITO). By varying Au and Ag thicknesses (10–50 nm) across a spectral range of 300–1500 nm, we generate spatial absorption maps and integrated power metrics from full-wave solutions to Maxwell’s equations. A multilayer perceptron models global absorption behavior with a mean absolute error (MAE) of 0.0953, while a convolutional neural network predicts spatial absorption distributions with an MAE of 0.0101. SHapley Additive exPlanations identify plasmonic layer thickness and excitation wavelength as dominant contributors to absorption, which peaks between 450 and 850 nm. Gold demonstrates broader and more sustained absorption compared to silver, although both metals exhibit reduced efficiency outside the resonance window. The integrated FDTD–ML framework accelerates plasmonic design while maintaining physical interpretability and predictive accuracy, enabling efficient exploration of tunable optical responses in multilayer nanophotonic systems for applications in optical sensing, photovoltaics, and device engineering.</div></div>","PeriodicalId":50941,"journal":{"name":"Advanced Engineering Informatics","volume":"71 ","pages":"Article 104311"},"PeriodicalIF":9.9,"publicationDate":"2026-01-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145926900","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 universal domain adaptive fault diagnosis method for ground source heat pump units with unknown fault awareness capability 具有未知故障感知能力的地源热泵机组通用域自适应故障诊断方法

IF 9.9 1区工程技术 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Advanced Engineering Informatics

Pub Date : 2026-01-09 DOI: 10.1016/j.aei.2025.104308

Shouqi Wang , Xuejin Gao , Huayun Han , Huihui Gao , Yongsheng Qi , Huazheng Han , Huimin Cheng

The Universal Domain Adaptation (UniDA) method can realize cross-device transfer diagnosis of ground source heat pump (GSHP) units when the class space of the target domain is unknown. However, existing methods generally design models on the premise that unknown faults exist in the target domain. When there are actually no unknown faults, the models often misclassify a large number of samples as unknown fault classes, thereby significantly reducing the diagnostic accuracy. Therefore, this paper proposes a UniDA method with Unknown Fault Awareness Capability (UFAC), which addresses the transfer challenge of inconsistent class spaces in GSHP units from a new perspective. First, the network trained on the source domain is used to perform feature extraction on the source domain and the target domain data, compute intra-class compactness and inter-class distance, and determine whether unknown fault classes exist in the target domain by combining One-sided T-test (OST) and Kernel Density Estimation (KDE). Subsequently, corresponding diagnostic models are constructed according to the judgment results, and the Balanced Adversarial Alignment (BAA) mechanism is introduced during training to achieve balanced cross-domain category distribution, unifying the model into the closed-set and open-set frameworks and improving cross-domain diagnostic efficiency. Experimental results show that this method achieves an average accuracy of 86.67 % in cross-device diagnosis of GSHP units, with performance significantly superior to existing methods, verifying its practicality and engineering application prospects under complex transfer environments.

通用域自适应（UniDA）方法可以在目标域类空间未知的情况下实现地源热泵机组的跨设备转移诊断。然而，现有的方法通常以目标域存在未知故障为前提来设计模型。在实际不存在未知故障的情况下，模型往往会将大量样本误分类为未知故障类，从而大大降低了诊断准确率。因此，本文提出了一种具有未知故障感知能力（UFAC）的UniDA方法，从一个新的角度解决了地源热泵机组中类空间不一致的传输挑战。首先，利用源域训练好的网络对源域和目标域数据进行特征提取，计算类内紧密度和类间距离，结合单侧t检验（OST）和核密度估计（KDE）确定目标域是否存在未知故障类。随后，根据判断结果构建相应的诊断模型，并在训练过程中引入平衡对抗性对齐（BAA）机制，实现跨域类别均衡分布，将模型统一为闭集和开集框架，提高跨域诊断效率。实验结果表明，该方法在地源热泵机组跨设备诊断中的平均准确率达到86.67%，性能明显优于现有方法，验证了其在复杂传递环境下的实用性和工程应用前景。

{"title":"A universal domain adaptive fault diagnosis method for ground source heat pump units with unknown fault awareness capability","authors":"Shouqi Wang , Xuejin Gao , Huayun Han , Huihui Gao , Yongsheng Qi , Huazheng Han , Huimin Cheng","doi":"10.1016/j.aei.2025.104308","DOIUrl":"10.1016/j.aei.2025.104308","url":null,"abstract":"<div><div>The Universal Domain Adaptation (UniDA) method can realize cross-device transfer diagnosis of ground source heat pump (GSHP) units when the class space of the target domain is unknown. However, existing methods generally design models on the premise that unknown faults exist in the target domain. When there are actually no unknown faults, the models often misclassify a large number of samples as unknown fault classes, thereby significantly reducing the diagnostic accuracy. Therefore, this paper proposes a UniDA method with Unknown Fault Awareness Capability (UFAC), which addresses the transfer challenge of inconsistent class spaces in GSHP units from a new perspective. First, the network trained on the source domain is used to perform feature extraction on the source domain and the target domain data, compute intra-class compactness and inter-class distance, and determine whether unknown fault classes exist in the target domain by combining One-sided T-test (OST) and Kernel Density Estimation (KDE). Subsequently, corresponding diagnostic models are constructed according to the judgment results, and the Balanced Adversarial Alignment (BAA) mechanism is introduced during training to achieve balanced cross-domain category distribution, unifying the model into the closed-set and open-set frameworks and improving cross-domain diagnostic efficiency. Experimental results show that this method achieves an average accuracy of 86.67 % in cross-device diagnosis of GSHP units, with performance significantly superior to existing methods, verifying its practicality and engineering application prospects under complex transfer environments.</div></div>","PeriodicalId":50941,"journal":{"name":"Advanced Engineering Informatics","volume":"71 ","pages":"Article 104308"},"PeriodicalIF":9.9,"publicationDate":"2026-01-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145926901","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 cross-working-condition prediction method for bearing remaining useful life based on SPW-SVDD health indicators and temporal-self -attention mechanism 基于SPW-SVDD健康指标和时间自关注机制的轴承剩余使用寿命交叉工况预测方法

IF 9.9 1区工程技术 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Advanced Engineering Informatics

Pub Date : 2026-01-09 DOI: 10.1016/j.aei.2026.104313

Zhao Yongxing, Ta Yuntian, Bi Ran, Tang Bo, Lu Zhengjie, Yan Yihong, Xie Jingsong, Guo Zhibin

This paper tackles the critical challenges in bearing Remaining Useful Life (RUL) prediction, including insecure health indicator construction, ambiguous degradation damage processes, and inadequate cross-working-condition generalizability. It presents a multi-feature HI relying on joint feature distribution modeling and a RUL prediction architecture based on temporal −self-attention (TSA) based Dual-branch Transfer Adversarial Network (TSADTAN). Firstly, a self-supervised probabilistic whitened support vector data description (SPW-SVDD) method is put forward to construct multi-feature health indicators with strong robustness. Secondly, a TSA mechanism is devised to boost the model’s capability to detect early minor faults and cumulative degradation features. Finally, a dual-branch adversarial transfer learning framework is built. By combining maximum mean discrepancy (MMD) and adversarial training through the domain comment feature (CF) branch and domain specific feature (SF) branch, stable feature alignment and knowledge transfer under cross-working conditions are achieved. Four cross-working-condition transfer prediction tasks are designed on two public bearing datasets. Experimental results show that the proposed method outperforms existing mainstream methods in three evaluation metrics, verifying its feasibility and effectiveness in practical cross-working-condition tasks.

本文解决了轴承剩余使用寿命（RUL）预测中的关键挑战，包括不安全的健康指标构建，模糊的退化损伤过程，以及不充分的跨工况通用性。提出了一种基于联合特征分布建模的多特征HI和基于时间自关注（TSA）的双分支转移对抗网络（TSADTAN）的RUL预测体系结构。首先，提出一种自监督概率白化支持向量数据描述（SPW-SVDD）方法，构建具有强鲁棒性的多特征健康指标；其次，设计了一种TSA机制来提高模型对早期小故障和累积退化特征的检测能力。最后，构建了一个双分支对抗迁移学习框架。通过领域评论特征（CF）分支和领域特定特征（SF）分支，将最大平均差异（MMD）和对抗训练相结合，实现了跨工况下稳定的特征对齐和知识转移。在两个公共轴承数据集上设计了4个跨工况转移预测任务。实验结果表明，该方法在三个评价指标上均优于现有主流方法，验证了该方法在实际跨工况任务中的可行性和有效性。

{"title":"A cross-working-condition prediction method for bearing remaining useful life based on SPW-SVDD health indicators and temporal-self -attention mechanism","authors":"Zhao Yongxing, Ta Yuntian, Bi Ran, Tang Bo, Lu Zhengjie, Yan Yihong, Xie Jingsong, Guo Zhibin","doi":"10.1016/j.aei.2026.104313","DOIUrl":"10.1016/j.aei.2026.104313","url":null,"abstract":"<div><div>This paper tackles the critical challenges in bearing Remaining Useful Life (RUL) prediction, including insecure health indicator construction, ambiguous degradation damage processes, and inadequate cross-working-condition generalizability. It presents a multi-feature HI relying on joint feature distribution modeling and a RUL prediction architecture based on temporal −self-attention (TSA) based Dual-branch Transfer Adversarial Network (TSADTAN). Firstly, a self-supervised probabilistic whitened support vector data description (SPW-SVDD) method is put forward to construct multi-feature health indicators with strong robustness. Secondly, a TSA mechanism is devised to boost the model’s capability to detect early minor faults and cumulative degradation features. Finally, a dual-branch adversarial transfer learning framework is built. By combining maximum mean discrepancy (MMD) and adversarial training through the domain comment feature (CF) branch and domain specific feature (SF) branch, stable feature alignment and knowledge transfer under cross-working conditions are achieved. Four cross-working-condition transfer prediction tasks are designed on two public bearing datasets. Experimental results show that the proposed method outperforms existing mainstream methods in three evaluation metrics, verifying its feasibility and effectiveness in practical cross-working-condition tasks.</div></div>","PeriodicalId":50941,"journal":{"name":"Advanced Engineering Informatics","volume":"71 ","pages":"Article 104313"},"PeriodicalIF":9.9,"publicationDate":"2026-01-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145926902","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

An improved tuna swarm optimization with dimension learning-based hunting for global optimization and real-world engineering applications 面向全局优化和实际工程应用的基于维度学习的改进金枪鱼群优化

IF 9.9 1区工程技术 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Advanced Engineering Informatics

Pub Date : 2026-01-08 DOI: 10.1016/j.aei.2025.104298

Eda Özkul

This study proposes an improved tuna swarm optimization algorithm (I-TSO) for solving global optimization and engineering design problems. However, despite its strong global search ability, tuna swarm optimization (TSO) suffers from trapping in local optima, having premature convergence, and the loss of diversity in the early stage. To eliminate these disadvantages and improve the original TSO, the proposed I-TSO algorithm uses a dimension learning-based hunting (DLH) strategy. DLH constructs a neighborhood for each tuna in the population and uses that information in the optimization process. Thus, it improves population diversity, provides a proper balance between exploration and exploitation, and prevents trapping into local optima. The performance of the proposed algorithm is evaluated on 23 classical benchmark functions, CEC-2017, CEC-2020, and CEC-2022 test suites, and compared it with eight other optimization algorithms. Comparative results demonstrate that I-TSO exhibits stable and effective optimization capabilities. Further, the Friedman test and Wilcoxon signed-rank test are conducted to statistically evaluate the performance of the proposed algorithm, and thus its superiority is statistically confirmed. Moreover, the applicability of the I-TSO in real-world problems is validated on eight engineering design problems. Consequently, the I-TSO algorithm is capable of solving both numerical and engineering design problems with its efficient and superior performance.

本文提出了一种改进的金枪鱼群优化算法（I-TSO），用于解决全局优化和工程设计问题。然而，金枪鱼群优化算法（TSO）虽然具有较强的全局搜索能力，但存在陷入局部最优、过早收敛、早期多样性丧失等问题。为了消除这些缺点并改进原TSO算法，本文提出的I-TSO算法采用了一种基于维学习的搜索（DLH）策略。DLH为种群中的每条金枪鱼构建一个邻域，并在优化过程中使用该信息。因此，它提高了种群多样性，在勘探和开采之间提供了适当的平衡，并防止陷入局部最优状态。在23个经典基准函数、CEC-2017、CEC-2020和CEC-2022测试套件上对该算法的性能进行了评估，并与其他8种优化算法进行了比较。对比结果表明，I-TSO具有稳定有效的优化能力。通过Friedman检验和Wilcoxon有符号秩检验对算法的性能进行统计评价，从而证实了算法的优越性。并通过8个工程设计问题验证了该方法在实际问题中的适用性。因此，I-TSO算法能够以其高效和优越的性能解决数值和工程设计问题。

{"title":"An improved tuna swarm optimization with dimension learning-based hunting for global optimization and real-world engineering applications","authors":"Eda Özkul","doi":"10.1016/j.aei.2025.104298","DOIUrl":"10.1016/j.aei.2025.104298","url":null,"abstract":"<div><div>This study proposes an improved tuna swarm optimization algorithm (I-TSO) for solving global optimization and engineering design problems. However, despite its strong global search ability, tuna swarm optimization (TSO) suffers from trapping in local optima, having premature convergence, and the loss of diversity in the early stage. To eliminate these disadvantages and improve the original TSO, the proposed I-TSO algorithm uses a dimension learning-based hunting (DLH) strategy. DLH constructs a neighborhood for each tuna in the population and uses that information in the optimization process. Thus, it improves population diversity, provides a proper balance between exploration and exploitation, and prevents trapping into local optima. The performance of the proposed algorithm is evaluated on 23 classical benchmark functions, CEC-2017, CEC-2020, and CEC-2022 test suites, and compared it with eight other optimization algorithms. Comparative results demonstrate that I-TSO exhibits stable and effective optimization capabilities. Further, the Friedman test and Wilcoxon signed-rank test are conducted to statistically evaluate the performance of the proposed algorithm, and thus its superiority is statistically confirmed. Moreover, the applicability of the I-TSO in real-world problems is validated on eight engineering design problems. Consequently, the I-TSO algorithm is capable of solving both numerical and engineering design problems with its efficient and superior performance.</div></div>","PeriodicalId":50941,"journal":{"name":"Advanced Engineering Informatics","volume":"71 ","pages":"Article 104298"},"PeriodicalIF":9.9,"publicationDate":"2026-01-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145926970","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

Large language models enable semantic-guided hierarchical games for intelligent battery coordination 大型语言模型支持语义引导的分层游戏，用于智能电池协调

IF 9.9 1区工程技术 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Advanced Engineering Informatics

Pub Date : 2026-01-08 DOI: 10.1016/j.aei.2026.104312

Yuntao Zou , Zihui Lin , Qianqi Zhang , Zhichun Liu , Zeling Xu

The battery energy consumption system of lunar exploration rovers, as mission-critical equipment, confronts severe challenges under extreme environmental constraints. However, existing modeling methods face fundamental dilemmas: dynamic uncertainty leads to highly ambiguous constraint boundaries, making it difficult for traditional mathematical languages to describe complex coupling relationships; even when mathematical representations are constructed, high-dimensional nonlinear optimization problems become computationally intractable, with existing algorithms unable to address complexity barriers and lacking interpretability. In response to these challenges, this paper innovatively proposes a hierarchical Stackelberg game optimization framework based on semantic embedding. This framework transcends traditional optimization paradigms by deeply integrating the cognitive intelligence of large language models with the mathematical precision of game theory: large language models acknowledge that overall behavior cannot be predicted from simple combinations of parts, processing fuzzy constraints and cross-domain knowledge integration through semantic understanding; the hierarchical structure of Stackelberg games naturally adapts to the hierarchical decision-making requirements of battery allocation, with multi-agent game frameworks effectively handling coordination and competition relationships between batteries. Through semantic embedding technology, natural language constraints are automatically transformed into mathematical objects comprehensible to game participants, with cognitive intelligence handling the “incomputable” complexity components while game theory ensures “provable” mathematical convergence, synergistically achieving the important paradigm transition from “perfect rationality” to “bounded rationality,” thereby providing a theoretically rigorous and practically viable unified solution for intelligent decision-making in mission-critical systems.

月球探测车电池能耗系统作为关键任务设备，在极端环境约束下面临严峻挑战。然而，现有的建模方法面临着根本性的困境：动态不确定性导致约束边界高度模糊，使得传统数学语言难以描述复杂的耦合关系；即使构建了数学表示，高维非线性优化问题在计算上也变得难以处理，现有算法无法解决复杂性障碍且缺乏可解释性。针对这些挑战，本文创新性地提出了一种基于语义嵌入的分层Stackelberg博弈优化框架。该框架超越了传统的优化范式，将大型语言模型的认知智能与博弈论的数学精度深度融合在一起：大型语言模型承认，整体行为不能通过简单的部件组合、模糊约束的处理以及通过语义理解进行跨领域知识整合来预测；Stackelberg博弈的分层结构自然适应了电池分配的分层决策要求，多智能体博弈框架有效处理了电池之间的协调与竞争关系。通过语义嵌入技术，自然语言约束自动转化为博弈参与者可以理解的数学对象，认知智能处理“不可计算”的复杂性成分，博弈论确保“可证明”的数学收敛，协同实现从“完全理性”到“有限理性”的重要范式转换。从而为关键任务系统的智能决策提供了理论严谨、实践可行的统一解决方案。

{"title":"Large language models enable semantic-guided hierarchical games for intelligent battery coordination","authors":"Yuntao Zou , Zihui Lin , Qianqi Zhang , Zhichun Liu , Zeling Xu","doi":"10.1016/j.aei.2026.104312","DOIUrl":"10.1016/j.aei.2026.104312","url":null,"abstract":"<div><div>The battery energy consumption system of lunar exploration rovers, as mission-critical equipment, confronts severe challenges under extreme environmental constraints. However, existing modeling methods face fundamental dilemmas: dynamic uncertainty leads to highly ambiguous constraint boundaries, making it difficult for traditional mathematical languages to describe complex coupling relationships; even when mathematical representations are constructed, high-dimensional nonlinear optimization problems become computationally intractable, with existing algorithms unable to address complexity barriers and lacking interpretability. In response to these challenges, this paper innovatively proposes a hierarchical Stackelberg game optimization framework based on semantic embedding. This framework transcends traditional optimization paradigms by deeply integrating the cognitive intelligence of large language models with the mathematical precision of game theory: large language models acknowledge that overall behavior cannot be predicted from simple combinations of parts, processing fuzzy constraints and cross-domain knowledge integration through semantic understanding; the hierarchical structure of Stackelberg games naturally adapts to the hierarchical decision-making requirements of battery allocation, with multi-agent game frameworks effectively handling coordination and competition relationships between batteries. Through semantic embedding technology, natural language constraints are automatically transformed into mathematical objects comprehensible to game participants, with cognitive intelligence handling the “incomputable” complexity components while game theory ensures “provable” mathematical convergence, synergistically achieving the important paradigm transition from “perfect rationality” to “bounded rationality,” thereby providing a theoretically rigorous and practically viable unified solution for intelligent decision-making in mission-critical systems.</div></div>","PeriodicalId":50941,"journal":{"name":"Advanced Engineering Informatics","volume":"71 ","pages":"Article 104312"},"PeriodicalIF":9.9,"publicationDate":"2026-01-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145926893","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

Energy management strategy for multisource power generation during hypersonic vehicle mode transition based on improved deep deterministic policy gradient 基于改进深度确定性策略梯度的高超声速飞行器模式转换多源发电能量管理策略

IF 9.9 1区工程技术 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Advanced Engineering Informatics

Pub Date : 2026-01-08 DOI: 10.1016/j.aei.2025.104293

Xingjian Jin, Fengying Zheng, Jingyang Zhang, Jiecheng Fu, Mengmeng Lv, Jian Lu, Si Gao

During the mode transition phase of hypersonic vehicles, stable electrical power supply is essential, while improper multisource power allocation tends to significantly increase fuel consumption or trigger engine safety-limit violations, further exacerbating the complexity of the power allocation problem. This paper aims to develop a deep reinforcement learning (DRL)-based energy management strategy to address these challenges and improve energy efficiency in advanced hypersonic propulsion systems. The strategy models the joint decision-making process of multisource power and engine fuel-flow rates distribution as a Markov decision process, and specifies a reward function that integrates cumulative additional fuel consumption, system demand constraints, and engine safety constraints. The double Q-learning with multi-replay buffer and adaptive soft update for deep deterministic policy gradient algorithm is introduced. First, it employs a state-partitioned multistage experience replay mechanism that organizes the experience buffer according to the sub-stages of the mode transition process. Second, an uncertainty-weighted double Q-network independently evaluates state-action pairs through two critic networks, fusing their outputs due to estimation uncertainty caused by engine state nonlinearities. Finally, a gradient-guided adaptive soft update adjusts the target network update rate, enabling smoother parameter updates amidst rapid engine and power generation transitions. The proposed energy management strategy effectively reduces fuel consumption, demonstrating strong policy adaptability, value estimation accuracy, and convergence stability. It achieved a fuel saving of 18.80% compared to the traditional demand-driven scheme. Relative to the basic DRL-based strategy, it improved the average reward growth by 17.9%, cumulative reward curve area by 23.5%, confidence interval convergence rate by 28.6%, and achieved a fuel saving of 4.6%.

在高超声速飞行器模态转换阶段，稳定的电力供应是必不可少的，而多源功率分配不当往往会显著增加燃油消耗或引发发动机安全极限违规，进一步加剧了功率分配问题的复杂性。本文旨在开发一种基于深度强化学习（DRL）的能量管理策略，以解决这些挑战并提高先进高超声速推进系统的能源效率。该策略将多源功率和发动机燃油流量分布的联合决策过程建模为马尔可夫决策过程，并指定了一个集成了累积额外燃油消耗、系统需求约束和发动机安全约束的奖励函数。介绍了深度确定性策略梯度算法中带有多重放缓冲和自适应软更新的双q学习算法。首先，采用状态划分的多阶段经验重放机制，根据模式转换过程的子阶段组织经验缓冲区。其次，不确定性加权双q网络通过两个批评网络独立评估状态-行为对，由于发动机状态非线性引起的估计不确定性，融合了它们的输出。最后，梯度引导的自适应软更新调整目标网络更新速率，在发动机和发电快速转换中实现更平滑的参数更新。所提出的能量管理策略有效地降低了燃料消耗，表现出较强的策略适应性、值估计准确性和收敛稳定性。与传统的需求驱动方案相比，它节省了18.80%的燃料。相对于基于drl的基本策略，平均奖励增长提高17.9%，累计奖励曲线面积提高23.5%，置信区间收敛率提高28.6%，节油4.6%。

{"title":"Energy management strategy for multisource power generation during hypersonic vehicle mode transition based on improved deep deterministic policy gradient","authors":"Xingjian Jin, Fengying Zheng, Jingyang Zhang, Jiecheng Fu, Mengmeng Lv, Jian Lu, Si Gao","doi":"10.1016/j.aei.2025.104293","DOIUrl":"10.1016/j.aei.2025.104293","url":null,"abstract":"<div><div>During the mode transition phase of hypersonic vehicles, stable electrical power supply is essential, while improper multisource power allocation tends to significantly increase fuel consumption or trigger engine safety-limit violations, further exacerbating the complexity of the power allocation problem. This paper aims to develop a deep reinforcement learning (DRL)-based energy management strategy to address these challenges and improve energy efficiency in advanced hypersonic propulsion systems. The strategy models the joint decision-making process of multisource power and engine fuel-flow rates distribution as a Markov decision process, and specifies a reward function that integrates cumulative additional fuel consumption, system demand constraints, and engine safety constraints. The double Q-learning with multi-replay buffer and adaptive soft update for deep deterministic policy gradient algorithm is introduced. First, it employs a state-partitioned multistage experience replay mechanism that organizes the experience buffer according to the sub-stages of the mode transition process. Second, an uncertainty-weighted double Q-network independently evaluates state-action pairs through two critic networks, fusing their outputs due to estimation uncertainty caused by engine state nonlinearities. Finally, a gradient-guided adaptive soft update adjusts the target network update rate, enabling smoother parameter updates amidst rapid engine and power generation transitions. The proposed energy management strategy effectively reduces fuel consumption, demonstrating strong policy adaptability, value estimation accuracy, and convergence stability. It achieved a fuel saving of 18.80% compared to the traditional demand-driven scheme. Relative to the basic DRL-based strategy, it improved the average reward growth by 17.9%, cumulative reward curve area by 23.5%, confidence interval convergence rate by 28.6%, and achieved a fuel saving of 4.6%.</div></div>","PeriodicalId":50941,"journal":{"name":"Advanced Engineering Informatics","volume":"71 ","pages":"Article 104293"},"PeriodicalIF":9.9,"publicationDate":"2026-01-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145926892","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

From laboratory to field: normal map-aided multimodal instance segmentation for blasting fragmentation analysis 从实验室到现场：法线图辅助爆破破片分析的多模态实例分割

IF 9.9 1区工程技术 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Advanced Engineering Informatics

Pub Date : 2026-01-08 DOI: 10.1016/j.aei.2026.104319

Yulin Wang , Xin Wang , Yudi Tang , Xu Dai , Jinming Dong , Guangyao Si

Particle size analysis of rock fragments plays a crucial role in mining engineering. However, traditional non-contact image-based methods typically rely solely on RGB images, which are highly sensitive to illumination changes, shadow interference, and fragment texture. This reliance limits the accuracy and generalization capability of conventional approaches and often necessitates retraining when applied across different scenes. To address these issues, this study introduces normal maps and provides a detailed analysis of their advantages in representing rock fragment features. Furthermore, we propose a multimodal instance segmentation framework named Adaptive Feature Recombination Network (AFRNet). AFRNet incorporates a modality effectiveness perception mechanism to adaptively guide the fusion process while suppressing interference from unreliable modalities. In addition, it employs a multi-scale attention fusion module to fully exploit and utilize the strength of each modality. This study systematically compares three fusion strategies—data-level, feature-level, and decision-level—and conducts experiments under various modality combinations. Experimental results demonstrate that incorporating normal maps significantly improves segmentation accuracy and enhances model robustness in degraded environments such as low illumination and shadow interference. Moreover, the model trained in a laboratory environment is directly transferred, without retraining, to a practical particle size analysis task at an actual mining site in Nanjing. The resulting particle size distribution curves exhibit a deviation of less than 10% compared with manually labeled results, validating the proposed method’s zero-cost transferability and engineering applicability.

岩屑粒度分析在采矿工程中具有重要的意义。然而，传统的基于非接触式图像的方法通常仅依赖于RGB图像，RGB图像对光照变化、阴影干扰和碎片纹理高度敏感。这种依赖限制了传统方法的准确性和泛化能力，并且在应用于不同场景时通常需要重新训练。为了解决这些问题，本研究引入了法线图，并详细分析了法线图在表示岩石碎片特征方面的优势。在此基础上，提出了一种多模态实例分割框架——自适应特征重组网络（AFRNet）。AFRNet采用模态有效性感知机制自适应引导融合过程，同时抑制不可靠模态的干扰。此外，采用多尺度注意力融合模块，充分挖掘和利用各模态的优势。本研究系统比较了数据级、特征级和决策级三种融合策略，并在不同的模态组合下进行了实验。实验结果表明，在低照度和阴影干扰等退化环境下，采用法线映射可以显著提高分割精度，增强模型的鲁棒性。此外，在实验室环境中训练的模型直接转移到南京实际采矿现场的实际粒度分析任务中，而无需再训练。与人工标记的结果相比，所得的粒径分布曲线偏差小于10%，验证了该方法的零成本可转移性和工程适用性。

{"title":"From laboratory to field: normal map-aided multimodal instance segmentation for blasting fragmentation analysis","authors":"Yulin Wang , Xin Wang , Yudi Tang , Xu Dai , Jinming Dong , Guangyao Si","doi":"10.1016/j.aei.2026.104319","DOIUrl":"10.1016/j.aei.2026.104319","url":null,"abstract":"<div><div>Particle size analysis of rock fragments plays a crucial role in mining engineering. However, traditional non-contact image-based methods typically rely solely on RGB images, which are highly sensitive to illumination changes, shadow interference, and fragment texture. This reliance limits the accuracy and generalization capability of conventional approaches and often necessitates retraining when applied across different scenes. To address these issues, this study introduces normal maps and provides a detailed analysis of their advantages in representing rock fragment features. Furthermore, we propose a multimodal instance segmentation framework named Adaptive Feature Recombination Network (AFRNet). AFRNet incorporates a modality effectiveness perception mechanism to adaptively guide the fusion process while suppressing interference from unreliable modalities. In addition, it employs a multi-scale attention fusion module to fully exploit and utilize the strength of each modality. This study systematically compares three fusion strategies—data-level, feature-level, and decision-level—and conducts experiments under various modality combinations. Experimental results demonstrate that incorporating normal maps significantly improves segmentation accuracy and enhances model robustness in degraded environments such as low illumination and shadow interference. Moreover, the model trained in a laboratory environment is directly transferred, without retraining, to a practical particle size analysis task at an actual mining site in Nanjing. The resulting particle size distribution curves exhibit a deviation of less than 10% compared with manually labeled results, validating the proposed method’s zero-cost transferability and engineering applicability.</div></div>","PeriodicalId":50941,"journal":{"name":"Advanced Engineering Informatics","volume":"71 ","pages":"Article 104319"},"PeriodicalIF":9.9,"publicationDate":"2026-01-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145926899","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

Quality-driven unmanned compaction system with non-invasive retrofitting on existing roller 以质量为导向的无人压实系统，在现有压路机上进行无创改造

IF 9.9 1区工程技术 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Advanced Engineering Informatics

Pub Date : 2026-01-08 DOI: 10.1016/j.aei.2025.104259

Jiale Li , Weitao Li , Jianmin Zhang , Guowei Ma , Xuefei Wang

Unmanned compaction (UC) enhances construction automation and efficiency over manual methods, with autonomous driving and operational control as critical technologies. However, practical application is limited by destructive retrofitting, insufficient algorithm adaptability for rollers, and a lack of integration with real-time compaction quality. This study presents a quality-driven unmanned compaction system with non-invasive hardware and a hierarchical control algorithm. It incorporates a two-layer adaptive particle swarm optimization (APSO) framework for dynamic path planning in under-compacted zones. A robust nonlinear model predictive control (MPC) path tracker is developed for stable trajectory execution. It integrates an extended Kalman filter (EKF) and a disturbance observer (DOB). Field tests demonstrate a 64% reduction in maximum lateral error and an increase in the final average compaction degree to 94.8%, compared to 86.7% achieved by manual operation. By achieving full-process, closed-loop quality control, this research provides a systematic solution for efficient, high-quality autonomous subgrade construction with significant value for engineering practice.

与人工方法相比，无人压实（UC）提高了施工自动化和效率，自动驾驶和操作控制是关键技术。然而，实际应用受到破坏性改造，算法对辊的适应性不足以及缺乏与实时压实质量的集成的限制。本研究提出了一种具有非侵入性硬件和分层控制算法的质量驱动的无人压实系统。该算法采用两层自适应粒子群优化（APSO）框架进行欠紧区动态路径规划。为了稳定的轨迹执行，提出了一种鲁棒非线性模型预测控制（MPC）路径跟踪器。它集成了扩展卡尔曼滤波器（EKF）和扰动观测器（DOB）。现场测试表明，与人工操作的86.7%相比，最大横向误差减少了64%，最终平均压实度增加了94.8%。通过实现全过程、闭环的质量控制，为高效、高质量的自主路基施工提供了系统解决方案，具有重要的工程实践价值。

{"title":"Quality-driven unmanned compaction system with non-invasive retrofitting on existing roller","authors":"Jiale Li , Weitao Li , Jianmin Zhang , Guowei Ma , Xuefei Wang","doi":"10.1016/j.aei.2025.104259","DOIUrl":"10.1016/j.aei.2025.104259","url":null,"abstract":"<div><div>Unmanned compaction (UC) enhances construction automation and efficiency over manual methods, with autonomous driving and operational control as critical technologies. However, practical application is limited by destructive retrofitting, insufficient algorithm adaptability for rollers, and a lack of integration with real-time compaction quality. This study presents a quality-driven unmanned compaction system with non-invasive hardware and a hierarchical control algorithm. It incorporates a two-layer adaptive particle swarm optimization (APSO) framework for dynamic path planning in under-compacted zones. A robust nonlinear model predictive control (MPC) path tracker is developed for stable trajectory execution. It integrates an extended Kalman filter (EKF) and a disturbance observer (DOB). Field tests demonstrate a 64% reduction in maximum lateral error and an increase in the final average compaction degree to 94.8%, compared to 86.7% achieved by manual operation. By achieving full-process, closed-loop quality control, this research provides a systematic solution for efficient, high-quality autonomous subgrade construction with significant value for engineering practice.</div></div>","PeriodicalId":50941,"journal":{"name":"Advanced Engineering Informatics","volume":"71 ","pages":"Article 104259"},"PeriodicalIF":9.9,"publicationDate":"2026-01-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145926971","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