Applied Soft Computing最新文献

{"title":"Enhanced N-BEATS for mid-term electricity demand forecasting","authors":"Mateusz Kasprzyk ,&nbsp;Paweł Pełka ,&nbsp;Boris N. Oreshkin ,&nbsp;Grzegorz Dudek","doi":"10.1016/j.asoc.2025.113575","DOIUrl":"10.1016/j.asoc.2025.113575","url":null,"abstract":"<div><div>This paper presents an enhanced N-BEATS model, N-BEATS*, for improved mid-term electricity load forecasting (MTLF). Building on the strengths of the original N-BEATS architecture, which excels in handling complex time series data without requiring preprocessing or domain-specific knowledge, N-BEATS* introduces two key modifications. (1) A novel loss function — combining pinball loss based on MAPE with normalized MSE, the new loss function allows for a more balanced approach by capturing both <span><math><msub><mrow><mi>L</mi></mrow><mrow><mn>1</mn></mrow></msub></math></span> and <span><math><msub><mrow><mi>L</mi></mrow><mrow><mn>2</mn></mrow></msub></math></span> loss terms. (2) A modified block architecture — the internal structure of the N-BEATS blocks is adjusted by introducing a destandardization component to harmonize the processing of different time series, leading to more efficient and less complex forecasting tasks. Evaluated on real-world monthly electricity consumption data from 35 European countries, N-BEATS* demonstrates superior performance compared to its predecessor and other established forecasting methods, including statistical, machine learning, and hybrid models. N-BEATS* achieves the lowest MAPE and RMSE, while also exhibiting the lowest dispersion in forecast errors. The source code is publicly available at Kasprzyk (2025).</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"182 ","pages":"Article 113575"},"PeriodicalIF":7.2,"publicationDate":"2025-07-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144694369","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":"An adaptive strategy quantum particle swarm optimization method based on intuitionistic fuzzy entropy and evolutionary game theory","authors":"Guan Zhou,&nbsp;Zihao Fang,&nbsp;Yingxin Hu,&nbsp;Jintao Chen,&nbsp;Jinyu Ren","doi":"10.1016/j.asoc.2025.113654","DOIUrl":"10.1016/j.asoc.2025.113654","url":null,"abstract":"<div><div>Since premature decline in population diversity is a vital problem in heuristic algorithm optimization, numerous methods enhancing global search ability have been developed to avoid local optimum. However, global exploration strategy diverts resources from exploitation, reducing optimization accuracy. To maintain exploration ability while ensuring accuracy, an adaptive strategy quantum particle swarm optimization method (ASQPSO) based on intuitionistic fuzzy entropy (IFE) and evolutionary game theory (EGT) is proposed in this paper. Firstly, IFE is introduced to quantify algorithm population diversity. Next, this paper proposes several strategies and develops an algorithm structure based on EGT to improve exploration and exploitation performance. Finally, comparison experiments are conducted to verify the performance of ASQPSO. Test results on 23 benchmark functions indicate that the proposed method has better comprehensive performance than the comparison algorithms. This paper researches a feasible way to adjust the diversity of the QPSO method quantitatively and provides a reference for its application in the heuristic algorithms.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"183 ","pages":"Article 113654"},"PeriodicalIF":7.2,"publicationDate":"2025-07-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144703043","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":"Safety guardian of intelligent transportation: Fisheye image based blind zone detection for Super Large Articulated Bus (SLAB)","authors":"Fujian Liang ,&nbsp;Yongzhao Han ,&nbsp;Jirui Wang ,&nbsp;Xin Wang ,&nbsp;Hongjie Tang ,&nbsp;Jiaoyi Wu ,&nbsp;Zutao Zhang","doi":"10.1016/j.asoc.2025.113660","DOIUrl":"10.1016/j.asoc.2025.113660","url":null,"abstract":"<div><div>The Super Large Articulated Buses (SLAB), as a complement to road traffic in big cities, brings great convenience to residents. However, due to its body length of more than 30 m and its unique driving characteristic on the left side of the road, blind zone detection during right turns has garnered significant attention. This paper proposes an intelligent method using fisheye images to address such issues. The proposed strategy is primarily divided into three steps. Firstly, fisheye cameras are mounted on the side of the SLAB’s body to capture fisheye images, and the dual longitude method is employed for distortion correction. Secondly, a vehicle detection method based on Single Shot Multibox Detector (SSD) is proposed, which combines Squeeze-and-Excitation (SE) attention mechanism, Feature Pyramid Network (FPN) and Multi-branch Dilation Block (MDB), called MDB-SSD. Through ablation experiments, the mean average precision (<em>mAP</em>) of this model is observed to increase by 5.31 % on the <em>BDD100k</em> dataset and 7.68 % on the <em>VOC</em> dataset when compared to the baselines. Specifically, the mAP of the MDB-SSD model reaches 40.13 % on the BDD100k dataset and 83.42 % on the VOC dataset, demonstrating significant improvement in detection accuracy. The detection of fisheye images exhibits good robustness, enhancing vehicle detection performance for the blind zone during right turns in SLAB. Finally, based on fisheye images, the proposed cross-longitude distance measurement method demonstrates an average detection error of 3 % for forward distance and 9.8 % for lateral distance, providing convenience for SLAB’s assisted driving. The main focus of this paper provides a solution for the safe operation of SLAB.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"182 ","pages":"Article 113660"},"PeriodicalIF":7.2,"publicationDate":"2025-07-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144694367","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":"Generative emulation and uncertainty quantification of geological CO2 storage with conditional diffusion models","authors":"Zhongzheng Wang ,&nbsp;Yuntian Chen ,&nbsp;Guodong Chen ,&nbsp;Qiang Zheng ,&nbsp;Tianhao Wu ,&nbsp;Dongxiao Zhang","doi":"10.1016/j.asoc.2025.113542","DOIUrl":"10.1016/j.asoc.2025.113542","url":null,"abstract":"<div><div>Carbon capture and storage (CCS) has emerged as a pivotal technology for reaching climate-neutrality targets. Safe and effective deployment of CCS requires reliable predictions of pressure buildup and CO<span><math><msub><mrow></mrow><mrow><mn>2</mn></mrow></msub></math></span> plume migration under geological uncertainties. However, traditional numerical simulations are limited by computational inefficiency, while machine learning methods face bottlenecks in predictive accuracy and uncertainty. Here we introduce a generative emulation framework named DiffMF for efficient prediction of multiphase flows in geological CO<span><math><msub><mrow></mrow><mrow><mn>2</mn></mrow></msub></math></span> storage. The framework treats flow prediction as conditional generation processes and employs cutting-edge diffusion models to produce the temporal–spatial evolution of pressure and CO<span><math><msub><mrow></mrow><mrow><mn>2</mn></mrow></msub></math></span> saturation fields under varying geological property conditions. Unlike existing approaches that focus primarily on point estimation, the probabilistic nature of DiffMF allows for generating multiple predictions that align with the statistics of the underlying dynamics, thereby facilitating effective quantification of predictive uncertainty. Comprehensive evaluations on diverse CO<span><math><msub><mrow></mrow><mrow><mn>2</mn></mrow></msub></math></span> storage cases show that DiffMF achieves up to 52.6% lower CO<span><math><msub><mrow></mrow><mrow><mn>2</mn></mrow></msub></math></span> saturation error compared to leading baseline models while maintaining high accuracy even under increased geological heterogeneity. Furthermore, we interpret the black-box model via visual analysis, providing insights into the generation process of DiffMF. Finally, the application to uncertainty quantification and propagation task for a field-scale storage system demonstrates that DiffMF yields statistics of the system responses in close agreement with those derived from high-fidelity simulations while executing 100 times faster, underscoring its promising potential in practical applications. The proposed generative emulation paradigm enables real-time prediction and probabilistic modeling that can foster informed decision-making for CCS deployment.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"182 ","pages":"Article 113542"},"PeriodicalIF":7.2,"publicationDate":"2025-07-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144671095","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":"Expansion-Trajectory Optimization (ETO): A dual-operator metaheuristic for balanced global and local search","authors":"Erik Cuevas ,&nbsp;Oscar A. González-Sánchez ,&nbsp;Francisco Orozco- Jiménez ,&nbsp;Daniel Zaldívar ,&nbsp;Alma Rodríguez-Vazquez ,&nbsp;Ram Sarkar","doi":"10.1016/j.asoc.2025.113642","DOIUrl":"10.1016/j.asoc.2025.113642","url":null,"abstract":"<div><div>Premature convergence remains a critical limitation in many metaheuristic algorithms, and is often caused by a rapid loss of population diversity as individuals become overly similar early in the search process. To address this challenge, this paper proposes a new metaphor-free algorithm called Expansion-Trajectory Optimization (ETO), which introduces a dual-operator framework designed to maintain diversity and enhance search performance. The ETO algorithm combines two complementary mechanisms: the expansion operator, which leverages collective information from multiple individuals to identify and explore promising regions in the search space; and the trajectory operator, which conducts a guided search following a Fibonacci spiral. This spiral-based path enables a smooth transition from broad exploration to focused exploitation, thereby ensuring a balanced and adaptive search process. The proposed approach was rigorously evaluated against several state-of-the-art metaheuristic algorithms, using a diverse set of benchmark functions. The experimental results confirm that ETO achieves superior performance in terms of both accuracy and robustness, demonstrating its effectiveness in overcoming early convergence and enhancing optimization outcomes.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"183 ","pages":"Article 113642"},"PeriodicalIF":7.2,"publicationDate":"2025-07-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144703045","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":"3D CAD model dynamic clustering based on inertial feature encoder","authors":"Fangwei Ning ,&nbsp;Zirui Li ,&nbsp;Jiaxing Lu ,&nbsp;Yixuan Wang ,&nbsp;Yanxia Niu ,&nbsp;Yan Shi","doi":"10.1016/j.asoc.2025.113627","DOIUrl":"10.1016/j.asoc.2025.113627","url":null,"abstract":"<div><div>The number of three-dimensional (3D) computer-aided design (CAD) models of mechanical parts in cyber manufacturing has experienced explosive growth. Classified CAD model shape knowledge based on induction is conducive to model retrieval, design reuse, and machining reuse. However, 3D CAD feature extraction primarily utilizes projected views, point clouds, voxels, and meshes for dimensionality reduction. Nonetheless, complex processes and high computational costs impede effective shape analysis. Traditional distance measures in data spaces or shallow linear embedded spaces are susceptible to errors when assessing similarity in data clusters. Furthermore, as the size of the database increases, data distribution may change in dynamic clustering, leading to data drift. This paper proposes an automatic unsupervised learning shape classification method based on deep embedding for 3D mechanical part CAD models. First, an inertial feature descriptor that effectively represents shape characteristics was established to extract the multidimensional moment of inertia of the 3D CAD model. Second, the inertial feature data space was nonlinearly mapped to a low-dimensional feature space, and the clustering accuracy was improved through the joint training of the encoder and clustering layers. Simultaneously, we revealed the influence of <em>eps</em> and <em>min samples</em> of Density-based Spatial Clustering of Applications with Noise (DBSCAN) algorithm on the clustering distribution of the CAD models. Third, adding new data can effectively achieve dynamic clustering based on the original clustering results. This paper explains the potential problems of fuzzy clustering boundaries that may arise from adding new data. Experimental data showed that the silhouette coefficient calculated by the proposed method is 0.78, and the normalized mutual information is 0.82, which has an excellent automatic classification effect.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"182 ","pages":"Article 113627"},"PeriodicalIF":7.2,"publicationDate":"2025-07-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144686230","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":"UL-Phys:Ultra-lightweight remote physiological measurement in facial videos based on unsupervised learning","authors":"Haibo Zhang ,&nbsp;Xu Wang ,&nbsp;Pan Dang ,&nbsp;Chaohui Ma ,&nbsp;Shuai Liu ,&nbsp;Zhuang Xiong ,&nbsp;Cheng Liu","doi":"10.1016/j.asoc.2025.113593","DOIUrl":"10.1016/j.asoc.2025.113593","url":null,"abstract":"<div><div>Remote photoplethysmography (rPPG) enables non-contact monitoring of vital signs using facial videos, but current supervised learning methods often rely on complex architectures and large annotated datasets, limiting their practicality in real-time and resource-constrained scenarios. This paper addresses these limitations by proposing UL-Phys, an ultra-lightweight self-supervised framework for rPPG signal estimation. From a research standpoint, we reformulate the rPPG task as a linear self-supervised reconstruction problem, introducing a novel frequency-constrained objective to extract inherent periodic information without requiring ground truth labels. The framework integrates a lightweight 3D spatiotemporal encoder-decoder network, and a neuroscience-inspired hybrid attention module to enhance pulsatile signal regions while suppressing noise. Experimental evaluations on PURE and UBFC-rPPG datasets demonstrate that UL-Phys achieves superior performance compared to existing supervised and self-supervised baselines, while significantly reducing model complexity and inference latency. Our method also shows strong generalization across datasets, highlighting the value of embedding physiological priors into lightweight, self-supervised architectures. These findings offer a promising direction for scalable and deployable rPPG systems in real-world settings.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"182 ","pages":"Article 113593"},"PeriodicalIF":7.2,"publicationDate":"2025-07-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144665518","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":"Efficient perturbation-aware distinguishing score for zero-shot neural architecture search","authors":"Junhao Huang ,&nbsp;Bing Xue ,&nbsp;Yanan Sun ,&nbsp;Mengjie Zhang ,&nbsp;Gary G. Yen","doi":"10.1016/j.asoc.2025.113447","DOIUrl":"10.1016/j.asoc.2025.113447","url":null,"abstract":"<div><div>Zero-cost proxies are under the spotlight of neural architecture search (NAS) lately, thanks to their low computational cost in predicting architecture performance in a training-free manner. The NASWOT score is one of the representative proxies that measures the architecture’s ability to distinguish inputs at the activation layers. However, obtaining such a score still requires considerable calculations on a large kernel matrix about input similarity. Moreover, the NASWOT score is relatively coarse-grained and provides a rough estimation of the architecture’s ability to distinguish general inputs. In this paper, to further reduce the computational complexity, we first propose a simplified NASWOT scoring term by relaxing its original matrix-based calculation into a vector-based one. More importantly, we develop a fine-grained perturbation-aware term to measure how well the architecture can distinguish between inputs and their perturbed counterparts. We propose a layer-wise score multiplication approach to combine these two scoring terms, deriving a new proxy, named efficient perturbation-aware distinguishing score (ePADS). Experiments on various NAS spaces and datasets show that ePADS consistently outperforms other zero-cost proxies in terms of both predictive reliability and efficiency. Particularly, ePADS achieves the highest ranking correlation among the advanced competitors (e.g., Kendall’s coefficient of 0.620 on NAS-Bench-201 with ImageNet-16-120 and 0.485 on NDS-ENAS), and ePADS-based random architecture search spends only 0.018 GPU days on DARTS-CIFAR to find networks with an average error rate of 2.64%.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"182 ","pages":"Article 113447"},"PeriodicalIF":7.2,"publicationDate":"2025-07-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144663627","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":"Cooperative co-evolutionary search for meta multigraph and graph neural architecture on heterogeneous information networks","authors":"Yang Liu ,&nbsp;Xiangyi Teng ,&nbsp;Jing Liu","doi":"10.1016/j.asoc.2025.113541","DOIUrl":"10.1016/j.asoc.2025.113541","url":null,"abstract":"<div><div>To model the rich semantic information on heterogeneous information networks (HINs), heterogeneous graph neural architecture search (HGNAS) has become a research hotspot, as it offers a promising automatic search technique for heterogeneous graph neural networks (HGNNs). However, there is no method that can simultaneously solve the meta multigraph and neural architecture search, which are the two core problems of HGNAS. In addition, existing HGNAS methods can only search for the meta graph or determine the number of edge types by setting a threshold hyperparameter, which has limited expression or is difficult to determine and significantly affects performance. In this paper, a cooperative co-evolutionary meta multigraph and graph neural architecture search method (called CCMG) on HINs is proposed. Specifically, CCMG first represents the meta multigraph and neural architecture by discrete encodings, and the number of network layers is variable. Second, whether an encoding of the architecture is meaningful or not is affected by the value of the encoding taken at the corresponding meta multigraph position and their search space sizes are not imbalanced. To cope with these situations, they are cooperatively and collaboratively optimized in the form of subproblems, facilitating group collaboration and information sharing. Finally, the effectiveness and superiority of the CCMG are verified on six datasets for node classification and recommendation tasks. Over the comparison HGNAS method, CCMG improves its performance on the two tasks by an average of 2.29% and 1.21%, respectively.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"182 ","pages":"Article 113541"},"PeriodicalIF":7.2,"publicationDate":"2025-07-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144654214","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":"Dual-branch fusion network with mutual learning for 12-lead electrocardiogram signal classification","authors":"Ke Ma ,&nbsp;Tao Zhang ,&nbsp;Hengyuan Zhang ,&nbsp;Wu Huang","doi":"10.1016/j.asoc.2025.113638","DOIUrl":"10.1016/j.asoc.2025.113638","url":null,"abstract":"<div><div>At present, the vast majority of methods use 12-lead electrocardiograms as a two-dimensional array as network input, and use deep neural networks to extract inter-lead correlation features. However, extracting intra-lead differential features is particularly important, as not every lead’s feature carries equal significance for classification. In this paper, we propose a dual-branch fusion network with 12-lead separation and combination, integrating the idea of mutual learning. The dual-branch network extracts differentiated and correlated features respectively and fuse them for classification. Each branch network is not only supervised by the ground truth but also referenced the learning experience of another branch network to further improve its classification ability. To address data imbalance, we introduced a category weighted binary focal loss to increase the attention of the network to the samples with few classes. We validated the proposed method on two publicly available multi-label datasets. Compared to the baseline model, our model has significantly improved in performance, demonstrating strong competitiveness and validating the effectiveness of our method. The experimental results show that our proposed method surpasses existing methods and achieves state-of-the-art performance. The method enables lightweight deployment on wearable devices, such as 12-lead ECG garments and smartwatches, facilitating real-time arrhythmia monitoring.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"182 ","pages":"Article 113638"},"PeriodicalIF":7.2,"publicationDate":"2025-07-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144685480","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}