Jinglong Wang , Yu Zhang , Changju Liu , Jiangtao Xu
{"title":"AHDPC: Adaptively hyperbolic density peak clustering","authors":"Jinglong Wang , Yu Zhang , Changju Liu , Jiangtao Xu","doi":"10.1016/j.eswa.2025.130065","DOIUrl":"10.1016/j.eswa.2025.130065","url":null,"abstract":"<div><div>Non-uniformly distributed datasets are common in real-world, and density peak clustering (DPC) methods are used on these datasets due to their superior clustering performance. However, existing DPC relies on linearly growing Euclidean distance, causing misleading similarity between points from different clusters and limiting the improvement of accuracy. To overcome this limitation, this study introduces an adaptive hyperbolic density peak clustering algorithm (AHDPC) by extending DPC into hyperbolic space. First, linear Euclidean distance is replaced with exponentially growing hyperbolic distance to enhance density difference between different points. Then, to overcome the misclassification of points at the junction of high-density and low-density regions and errors from extreme hyperbolic distance, a novel adaptive weighting strategy is proposed, it dynamically adjusts hyperbolic distance by building the trace of the global covariance matrix, the Euclidean norm, the maximum pairwise distance, and point-to-center deviation. Finally, an adaptively cutoff distance method based on a segmented search strategy is developed to eliminate manual tuning, and an exponential density function replaces the gaussian kernel to improve computational efficiency. AHDPC not only overcomes the deficiencies of Euclidean space but also mitigates the restrictive aspects of hyperbolic space. Extensive experiments on synthetic and real datasets, the olivetti faces dataset, and medical image datasets demonstrate that AHDPC outperforms state-of-the-art methods in clustering accuracy. Results also show that AHDPC produces more discriminative decision graph for identifying cluster centers and enhances the accuracy of categorisation of non-center points. The advantages of its robustness and adaptive weight in improving the clustering performance are also confirmed.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"299 ","pages":"Article 130065"},"PeriodicalIF":7.5,"publicationDate":"2025-10-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145364919","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Fixed-time bipartite flocking of perturbed networked UAV systems: A distributed optimization approach","authors":"Weihao Li , Mengji Shi , Lei Shi , Boxian Lin","doi":"10.1016/j.eswa.2025.129979","DOIUrl":"10.1016/j.eswa.2025.129979","url":null,"abstract":"<div><div>Flocking, inspired by the collective dynamics observed in biological swarms, exemplifies the emergence of self-organization and swarm intelligence through local agent interactions. Motivated by these principles, this paper investigates the fixed-time bipartite flocking control problem for networked unmanned aerial vehicle (UAV) systems under external disturbances from the perspective of distributed optimization. The proposed solution addresses key concerns in multi-agent coordination, including the emergence rate of flocking behavior, robustness against uncertainties, and performance optimization. A hierarchical distributed optimal control framework is developed, consisting of a distributed optimization layer and a trajectory tracking layer. Theoretically, the proposed control scheme ensures (i) convergence to the global optimum of the distributed optimization problem within a fixed time, and (ii) fixed-time emergence of bipartite flocking behavior characterized by subgroup cohesion and velocity alignment. The stability of the closed-loop system is rigorously established via a Lyapunov-based analysis method, which also guarantees robustness against dynamic disturbances. In addition, explicit upper bounds on the settling time for both layers are derived, allowing the trade-off between convergence speed and control effort to be tuned through parameter selection. Finally, numerical simulations together with real-world experiments are presented to validate the effectiveness and practical feasibility of the proposed fixed-time bipartite flocking control scheme.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"299 ","pages":"Article 129979"},"PeriodicalIF":7.5,"publicationDate":"2025-10-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145334033","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Xin Zhou, Guodong Ling, Jiayi Yu, Tian Zhou, Rui Wang
{"title":"Balanced multi-objective evolution algorithm for unmanned systems project scheduling with preventive maintenance and order grouping constraints","authors":"Xin Zhou, Guodong Ling, Jiayi Yu, Tian Zhou, Rui Wang","doi":"10.1016/j.eswa.2025.130006","DOIUrl":"10.1016/j.eswa.2025.130006","url":null,"abstract":"<div><div>With the growing deployment of unmanned systems in multi-platform operations, intelligent scheduling has become increasingly critical. These systems are typically organized into distributed mission clusters, executing sequences of mission-critical operations under resource and operational constraints. Inspired by the structural similarity between unmanned systems coordination and manufacturing workflows, this paper reformulates the scheduling problem of unmanned missions as a distributed permutation flowshop scheduling problem. Two domain-specific factors are incorporated into the model: preventive maintenance and order grouping, with the objective of minimizing total tardiness and makespan. To address this problem, a balanced multi-objective evolution algorithm (BMOEA) is proposed. Initially, two improved heuristic algorithms based on NEH2 are used to balance the quality and diversity of initial solutions. Then, four target-specific operators and two crossover operators are designed to improve the search efficiency of the algorithm. Next, three criteria are developed to balance local and global search: a classification-based operator selection criterion, which dynamically adjusts the search direction of operators to optimize local search; a non-periodic evaluation criterion based on Kernel Density Estimation and a non-dominated solution threshold criterion, which accurately determines the timing for switching to global search. These criteria balance exploration and exploitation, allowing the algorithm to optimize both convergence speed and population diversity, expand the feasible domain, and steadily approach the Pareto front. Finally, the experimental results reveal that BMOEA delivers superior performance compared to the most advanced algorithms available.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"299 ","pages":"Article 130006"},"PeriodicalIF":7.5,"publicationDate":"2025-10-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145334026","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Chengcheng Xu, Tong Han, Tianfeng Wang, Xiao Han, Zhisong Pan
{"title":"Semantic-aware contrastive learning for graph classification","authors":"Chengcheng Xu, Tong Han, Tianfeng Wang, Xiao Han, Zhisong Pan","doi":"10.1016/j.eswa.2025.129976","DOIUrl":"10.1016/j.eswa.2025.129976","url":null,"abstract":"<div><div>Currently, the performance of vanilla graph-level contrastive learning is limited by traditional data augmentation strategies and size imbalance. First, adding noise in the embedding space for data augmentation may cause the graph-level representation to exceed the class decision boundary and alter its original semantic information. Second, transferring information from head graphs to tail graphs to alleviate size imbalance without distinguishing the semantic information of the graphs may lead to suboptimal model performance. To address these issues, we propose a semantic-aware graph-level contrastive learning method named SAGCL. Specifically, SAGCL achieves controllable data augmentation by adjusting the closeness between class center features and augmented features, preserving the inherent structure and semantic information of the graph. Meanwhile, the intra-cluster variance is used as a regularization term to maintain the uniformity of the feature distribution. In addition, SAGCL employs a confidence-weighted approach to obtain the semantic prototypes of head graphs and tail graphs within each cluster. Then, the rich semantic information from the head graphs is transferred to the tail graphs, effectively enhancing the model’s ability to distinguish tail graphs. Experiments on graph classification tasks on eight imbalanced datasets demonstrate that SAGCL outperforms existing state-of-the-art methods.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"299 ","pages":"Article 129976"},"PeriodicalIF":7.5,"publicationDate":"2025-10-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145364844","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Xiaomei Zhang, Min Deng, Jiwei Hu, Xiao Huang, Qiwen Jin
{"title":"CrysFormer++: Dual-phase refinement learning for transparent object depth estimation","authors":"Xiaomei Zhang, Min Deng, Jiwei Hu, Xiao Huang, Qiwen Jin","doi":"10.1016/j.eswa.2025.130043","DOIUrl":"10.1016/j.eswa.2025.130043","url":null,"abstract":"<div><div>Transparent object depth estimation is a critical yet challenging task in robotic perception, particularly in grasping applications for industrial automation and human-robot interaction. Due to the high transmittance of visible light in transparent materials, depth sensors often suffer from severe depth measurement errors, leading to inaccuracies in grasp planning and object manipulation. To address this issue, we propose a Mamba-Transformer hybrid encoding framework (CrysFormer++) for robust depth estimation of transparent objects. The model integrates VMamba to efficiently model global long-range dependencies and leverages Swin Transformer to capture fine-grained local features. In addition, we have developed a self-supervised confidence learning framework that generates pixel-wise reliability maps through photometric consistency constraints, and realizes adaptive fusion of raw depth measurements and network predictions via physics-informed spatial weighting. Meanwhile, we have designed a novel loss function to enhance the accuracy and robustness of depth prediction. Extensive experiments conducted on the TransCG and ClearGrasp datasets validate that CrysFormer++ achieves superior performance compared to existing state-of-the-art approaches, in terms of both visual quality and quantitative metrics. The results validate the effectiveness of CrysFormer++ in handling complex backgrounds, providing a high-precision depth perception solution for robotic grasping of transparent objects.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"299 ","pages":"Article 130043"},"PeriodicalIF":7.5,"publicationDate":"2025-10-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145364845","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"YOLO-DLA: A YOLO-based unified framework for multi-scale document layout analysis","authors":"Haoyan Qi, Xinyang Meng, Zhijuan Du","doi":"10.1016/j.eswa.2025.129981","DOIUrl":"10.1016/j.eswa.2025.129981","url":null,"abstract":"<div><div>Document layout analysis (DLA) serves as a cornerstone of modern information systems, enabling efficient data extraction and structured knowledge organization. However, processing multi-scale layout documents has remained a bottleneck in the development of universal DLA frameworks. To address this limitation, we introduce the first multi-scale DLA dataset and propose a novel YOLO-based detection framework. Specifically: (1) To address the lack of fine-grained layout annotations in existing datasets, we construct <strong>AcadLayout</strong>, a specialized dataset for scientific documents with 13 layout element types (e.g., multi-level headings, formula, figure caption). (2) To address the challenges of multi-scale feature extraction, particularly for micro-scale elements, we innovatively incorporate the KWConv dynamic convolution method. (3) To achieve robust feature fusion across scales, we propose the PRDM-neck module, which uniquely integrates axial attention with multi-scale context aggregation. (4) To address scale imbalance in scientific documents, we propose a scale-aware curriculum learning strategy that progressively trains models from macro- to micro-scale elements (macro<span><math><mo>→</mo></math></span>medium<span><math><mo>→</mo></math></span>micro), effectively balancing detection performance across all scales.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"299 ","pages":"Article 129981"},"PeriodicalIF":7.5,"publicationDate":"2025-10-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145364846","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Towards enhancing prototypes driven by graph convolutional network for domain adaptation","authors":"Ba Hung Ngo , Tae Jong Choi , Sung In Cho","doi":"10.1016/j.eswa.2025.130010","DOIUrl":"10.1016/j.eswa.2025.130010","url":null,"abstract":"<div><div>Domain adaptation (DA) is essential for transferring knowledge across domains with differing distributions, yet challenges like domain shifts and scarce labeled data limit performance. Prototype-based methods show promise on the DA task. This work introduces a prototype-based method, termed enhanced prototypical network (EnPro), for unsupervised domain adaptation (UDA) and semi-supervised domain adaptation (SSDA) settings with consistent architecture and training. We provide a theoretical analysis dividing the DA mapping space into <em>consensus, vicinal</em>, and <em>vulnerable</em> spaces. This improves classification by expanding the <em>consensus</em> and <em>vicinal</em> spaces while reducing the <em>vulnerable</em> space. To achieve this, we use a graph convolutional network (GCN) to increase <em>labeled</em> target samples through reliable pseudo-labels and enhanced prototypes. Experiments on UDA and SSDA benchmark datasets demonstrate state-of-the-art performance.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"299 ","pages":"Article 130010"},"PeriodicalIF":7.5,"publicationDate":"2025-10-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145364917","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Junchang Zhang , Yucai Shi , Hong Chen , Qing Wang , Hai Huang
{"title":"Low-light image enhancement integrated semantic aware guidance","authors":"Junchang Zhang , Yucai Shi , Hong Chen , Qing Wang , Hai Huang","doi":"10.1016/j.eswa.2025.130045","DOIUrl":"10.1016/j.eswa.2025.130045","url":null,"abstract":"<div><div>Simply increasing brightness is not an optimal solution for low-light image enhancement, as excessive adjustment often leads to overexposure, amplified noise, and color distortion. To address this issue, we propose a dual-branch semantic-aware guided enhancement network. One branch focuses on natural brightness adjustment, while the other conducts coarse semantic segmentation to guide region-specific enhancement. Indoor scenes are segmented into foreground and background, whereas outdoor scenes are divided into sky, foreground, and ground, with each region enhanced using tailored strategies. The backbone employs lightweight inverted residual convolutional blocks with attention mechanisms, and spatial-positional encoding is incorporated to inject absolute positional cues, thereby improving the understanding of image structures and spatial relationships. Extensive experiments on both the Outdoor-Synthetic dataset (synthesized from CamVid and Cityscapes) and the Indoor-LLRGDB_Real low-light dataset demonstrate that our method consistently surpasses state-of-the-art approaches in both qualitative and quantitative evaluations, achieving 33.448/0.977/0.141 (PSNR/SSIM/LPIPS) on Outdoor-Synthetic dataset and 17.939/0.993/0.259 on Indoor-LLRGDB_Real dataset. Furthermore, no-reference image quality assessments confirm the naturalness and realism of the enhanced results. Our code and corresponding database can be obtained at <span><span>https://github.com/zhangjunchang2023/LLIE-SAG</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"299 ","pages":"Article 130045"},"PeriodicalIF":7.5,"publicationDate":"2025-10-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145364983","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Effective hybrid branch-and-cut algorithm for the inventory routing problem with open vehicle routing constraints","authors":"Nai-Kang Yu , Bin Qian , Rong Hu , Jian-Bo Yang","doi":"10.1016/j.eswa.2025.129905","DOIUrl":"10.1016/j.eswa.2025.129905","url":null,"abstract":"<div><div>This study considers a kind of inventory routing problem, which jointly optimizes the open vehicle routing decision and the inventory replenishment in a real-world distribution scenario. That is, the considered problem is an integrated optimization problem with two coupled subproblems (IOP_TCSP), i.e., the open vehicle routing problem (OVRP) and the inventory replenishment problem. The criterion is to minimize the total logistics and inventory costs under multiple periods. The IOP_TCSP is modelled as a mixed integer programming problem, and then a hybrid branch-and-cut algorithm combining novel Lagrangian heuristic approach and valid inequalities (HB&C_NLHAVI) is devised to deal with it. Test results on 72 instances with different scales demonstrate that the devised HB&C is more effective than the commercial solver Gurobi. Specifically, the HB&C can obviously reduce optimality gaps for many instances within the similar or less running time, and it can reduce the optimality gaps by 12–24% within only 60–70% of Gurobi’s running time for almost all large-scale instances.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"299 ","pages":"Article 129905"},"PeriodicalIF":7.5,"publicationDate":"2025-10-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145334113","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Integrating BiLSTM and CNN for predicting user mobility from geotagged social media posts","authors":"Zhao Yu , Zohre Moradi","doi":"10.1016/j.eswa.2025.130004","DOIUrl":"10.1016/j.eswa.2025.130004","url":null,"abstract":"<div><div>Recommender systems for social media platforms face significant challenges in accurately capturing user preferences from unstructured geotagged data. This research introduces a hybrid recommendation system leveraging Convolutional neural networks (CNNs) paired with Bidirectional Long Short-Term Memory (BiLSTM) networks to improve the prediction of user mobility and preferences. The proposed model calculates user similarity by analyzing opinions and preferences extracted from social media posts, combining CNN strength in feature extraction with BiLSTM ability to capture users dependencies. By incorporating demographic data, the system addresses the cold-start issue and improves recommendation accuracy by utilizing contextual information. Experimental results using datasets from Yelp and Flickr demonstrate significant advancements in RMSE, F-Score, MAP, and NDCG metrics. These findings highlight the effectiveness of the CNN-BiLSTM hybrid approach in generating personalized, sentiment-aware, and contextually rich recommendations on social media platforms.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"299 ","pages":"Article 130004"},"PeriodicalIF":7.5,"publicationDate":"2025-10-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145364984","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}