Applied Soft Computing最新文献

{"title":"Detail-aware semantic segmentation network for brain tumor MRI images combining multi-frequency directional filtering and lifting wavelets","authors":"Xin Hua ,&nbsp;Zhijiang Du ,&nbsp;Hongjian Yu ,&nbsp;Zibo Li ,&nbsp;Qiaohui Lu ,&nbsp;Hui Zhao","doi":"10.1016/j.asoc.2025.113969","DOIUrl":"10.1016/j.asoc.2025.113969","url":null,"abstract":"<div><div>Brain tumors segmentation in Magnetic Resonance Imaging (MRI) images poses significant challenges owing to the uncertain location and size of the tumors, the difficulty in describing their boundaries, and the fuzzy demarcation of diseased tissues. Although U-Net and its recent variants have emerged as leading models for semantic segmentation in medical imaging, they still face structural limitations. These limitations cause the erosion of detail information during downsampling and poor performance in segmenting small lesions when handling targets of varying sizes, indicating a lack of detail handling capability. To counteract these issues, we designed a segmentation model that enhances detail features using frequency information. To reduce the loss of feature information during downsampling, we developed a downsampling module based on lifting wavelets. By lifting wavelets to group and integrate features according to frequency from high to low, we reduce feature resolution while enhancing information transmission and minimizing feature information loss. In our designed multi-frequency directional filtering edge feature extraction module, we extract low-frequency and high-frequency features and construct a dual-channel multi-directional filtering combination. This combination extracts directional information from low-frequency and high-frequency features separately, increasing the multi-angle directional information of the features and enriching the detailed information such as direction and position within the features. On the BraTS2018, BraTS2020, and BraTS2024 brain tumor datasets, our model demonstrated optimal results compared to 14 other advanced models. The average Dice Similarity Coefficients are 78.48 %, 79.80 %, and 74.35 %, while the 95th percentile Hausdorff Distances are 5.75, 6.60, and 7.72. Our code link is <span><span>https://github.com/Eric-H8/BraTS_Seg_Model</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"185 ","pages":"Article 113969"},"PeriodicalIF":6.6,"publicationDate":"2025-09-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145097646","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":"Industrial large models as carriers for malicious payloads: A fast and robust approach","authors":"Yi Yuan ,&nbsp;Ruijun Deng ,&nbsp;Zhihui Lu ,&nbsp;Patrick C.K. Hung","doi":"10.1016/j.asoc.2025.113967","DOIUrl":"10.1016/j.asoc.2025.113967","url":null,"abstract":"<div><div>As artificial intelligence (AI) technologies are increasingly deployed in industrial domains, the use of pre-trained industrial large models (ILMs) has become widespread due to their cost-effectiveness and high performance across complex tasks. However, this growing reliance introduces new cybersecurity threats, particularly concerning the integrity of model parameters. Attackers are increasingly targeting AI models, exploiting vulnerabilities in the software supply chain, providing a covert means of executing novel cyberattacks, and posing significant security risks. Although antivirus software and intrusion detection systems are effective in protecting systems, the evolving nature of attack strategies, particularly the use of widely available AI models as carriers for malicious payloads, poses increasingly sophisticated security threats. Most existing embedding techniques struggle in ILM application scenarios, where payloads are inefficiently embedded and extracted, and are easily disrupted by fine-tuning. Meanwhile, the few robustness techniques always face limitations in areas such as low efficiency, model performance degradation, and the challenge of achieving a balance between efficiency and robustness. In this paper, we introduce <strong>FREEZER</strong>: Fast Redundant Exponent Embedding with Robustness (Robustness refers to the ability to preserve the embedded payload after full fine-tuning, while fast speed denotes its substantially faster payload embedding and extraction compared to prior approaches), a framework for significantly improving efficiency while maintaining robustness during the injection of malicious payloads into ILMs. FREEZER effectively addresses the challenge of extensive bit errors—defined as the bitwise discrepancy between the originally embedded malicious payload and the payload recovered by the prescribed extractor after full-parameter fine-tuning. Moreover, the infected models obtained using FREEZER exhibit no significant performance degradation. Experimental results show that FREEZER achieves a 20x faster injection speed and a 240x faster extraction speed compared to the current state-of-the-art (SOTA) method while maintaining high robustness. FREEZER raises awareness of this emerging threat and inspires the development of novel defenses against new forms of cyberattacks.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"185 ","pages":"Article 113967"},"PeriodicalIF":6.6,"publicationDate":"2025-09-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145268650","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 driving route planning based on asymmetric multi-agent path planning problem with limited service area constraints","authors":"Ali Maktabifard,&nbsp;Dávid Földes","doi":"10.1016/j.asoc.2025.113959","DOIUrl":"10.1016/j.asoc.2025.113959","url":null,"abstract":"<div><div>Mobility and logistics service providers operating a vehicle fleet and serving several targets with each vehicle are faced with the challenge of planning efficient and cooperative driving routes while considering all the vehicles. To address this challenge, we present a novel driving route planning method based on an Asymmetric Multi-Agent Path Planning (AMAPP) problem, a variation of the classical Multiple Traveling Salesmen Problem (MTSP). Given a team of <span><math><mi>m</mi></math></span> agents (vehicles) that must visit <span><math><mi>n</mi></math></span> targets located in a real road network, a set of <span><math><mi>m</mi></math></span> optimized open paths (the <em>Plan</em>) must be found such that each target is visited exactly once. The optimization objective is to minimize the driving distance of the path with the longest driving distance in the Plan (a Min-Max problem) so that this cooperative operation can be completed in the least amount of time. In several cases, the service area for each agent is limited (e.g., certain districts in a city). To simulate this real-world condition, two additional constraints are applied: the maximum geodetic range for each agent, and the maximum spatial range of targets for each agent. An easy-to-apply Genetic Algorithm (GA) with two novel initialization methods is presented to solve this route planning problem. In order to validate the developed route planning method and demonstrate its applicability, the method was tested in a real-world test case where it showed a decent performance. The results show that the applied limited service area constraints not only decrease the average driving distance of the longest route in the route plan, but also reduce the average runtime of the developed solution. Additionally, the performance of the proposed GA was benchmarked using 26 problems based on the TSPLIB instances, where the proposed GA achieved new Best Known Solution (BKS) values for 19 benchmark problems and a result equal to the BKS value for another two problems, demonstrating its superiority and robustness. The developed method can be used for route planning of the mobility and logistics services requiring multiple destinations to be reached by a fleet of vehicles, such as group ride-sharing and last-mile delivery.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"185 ","pages":"Article 113959"},"PeriodicalIF":6.6,"publicationDate":"2025-09-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145221559","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":"A federated learning-based method for personalized manufacturing service recommendation with collaborative relationships","authors":"Lei Wang ,&nbsp;Jun Wang ,&nbsp;Feng Xiang ,&nbsp;Tongshun Li ,&nbsp;Yang Xu ,&nbsp;Yibing Li","doi":"10.1016/j.asoc.2025.113940","DOIUrl":"10.1016/j.asoc.2025.113940","url":null,"abstract":"<div><div>In the industrial Internet environment, the increasing complexity of manufacturing tasks has rendered them no longer accomplishable by independent manufacturing services. Meanwhile, current recommendation systems predominantly face challenges in maintaining data privacy and security during client parameter exchanges. To address these issues, this paper proposes CoFedSVD+ +, a federated learning-based method for personalized manufacturing service recommendation that integrates an enhanced SVD+ + algorithm with homomorphic encryption. First, we devise an enhanced similarity calculation method to analyze collaborative relationships among manufacturing services. Second, we implement a homomorphic encryption protocol within the federated learning framework to resolve data isolation challenges. Third, the improved SVD+ + algorithm is employed to capture implicit feedback information and predict missing Quality of Service (QoS) metrics. Fourth, a Top-N service composition recommendation list is generated through synergistic analysis of collaborative relationships and QoS predictions. Finally, we validate our approach using real-world case data from an industrial Internet platform. Experimental comparisons with existing recommendation algorithms demonstrate superior recommendation effectiveness of the proposed method.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"185 ","pages":"Article 113940"},"PeriodicalIF":6.6,"publicationDate":"2025-09-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145119756","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":"A collaborative competition multitasking framework for constrained multi-objective optimization","authors":"Xinyu Feng ,&nbsp;Qianlong Dang ,&nbsp;Xiaochuan Gao ,&nbsp;Yanghui Wu ,&nbsp;Lifei Zheng","doi":"10.1016/j.asoc.2025.113900","DOIUrl":"10.1016/j.asoc.2025.113900","url":null,"abstract":"<div><div>Constrained multi-objective optimization problems (CMOPs) are common in the real world. Constrained multi-objective evolutionary algorithms (CMOEAs) based on evolutionary multi-tasking show excellent performance in solving CMOPs. However, not all tasks can find useful information during the process of evolution, which inevitably results in a waste of computing resources. In this paper, a CMOEA based on collaborative competition multitasking (TCCMT) is proposed, in which two auxiliary tasks are constructed to co-evolve with the main task in a collaborative competition manner. During the process of evolution, only the dominant auxiliary task is selected to help the main task evolve, which reduces the resource consumption to evolve the invalid tasks. Meanwhile, the evolutionary process is divided into three stages in order to balance exploration and exploitation. The auxiliary tasks customize the constrained adaptive regression strategy and double angle enhancement strategy respectively to improve the ability to solve different problems. Compared with the nine most advanced CMOEAs on 33 benchmark problems and 7 real-world engineering problems, the Friedman test results show that TCCMT achieves the best rank on all test problems and exhibits a statistically significant difference.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"185 ","pages":"Article 113900"},"PeriodicalIF":6.6,"publicationDate":"2025-09-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145221560","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":"Representing online reviews using interval type-2 fuzzy Z-numbers for ranking energy-saving appliances","authors":"Yue Xiao ,&nbsp;Ming Li ,&nbsp;Ying Li ,&nbsp;Hongde Liu","doi":"10.1016/j.asoc.2025.113961","DOIUrl":"10.1016/j.asoc.2025.113961","url":null,"abstract":"<div><div>As e-commerce develops and the green consumption concept gains popularity, consumers are increasingly inclined to purchase energy-saving home appliances through e-commerce platforms. However, they often face technical complexities related to specialized energy-saving attributes and an overwhelming number of online reviews. To address these challenges, we have integrated energy-saving features into our online review analysis, incorporating weight calculations. Notably, we propose and prove a method that transforms online review information into interval type-2 fuzzy Z-numbers (IT2FZNs), which comprehensively represent the information to support product ranking. First, based on the energy-saving attributes of energy-saving appliances and their online review data, we extract energy-saving features and online review features, respectively. We then use a combination of TF-IDF-based text mining and BWM-based expert evaluation to determine the weight of each feature. Next, we convert the energy-saving feature data into IT2FZNs according to specific rules. The online review feature data is converted into interval type-2 fuzzy sets (IT2Fs) by considering the sentiment classification results and the accuracy and robustness of the model, and is further combined with the reliability of online reviews to form IT2FZNs. Finally, the alternative products are ranked based on the constructed decision matrix, and the final ranking results are determined. The method's effectiveness and practicality have been demonstrated using real data from energy-saving refrigerators on the JingDong (JD.com) platform, and its robustness and superiority have been further substantiated through comparative experiments.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"185 ","pages":"Article 113961"},"PeriodicalIF":6.6,"publicationDate":"2025-09-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145183686","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":"FFNN: Fractional order basis function multi-step neural network method for fractional partial differential equations","authors":"Jianke Zhang,&nbsp;Xudong Tian,&nbsp;Chang Zhou","doi":"10.1016/j.asoc.2025.113907","DOIUrl":"10.1016/j.asoc.2025.113907","url":null,"abstract":"<div><div>With the advancement in artificial intelligence technology, the increasing number of researchers utilize it to address complex equations in ocean engineering. So the technology of artificial intelligence has become a practical area of research. In this paper, we design a novel method to solve the fractional order long water wave equation, which is called the fractional order basis function multi-step neural network. Firstly, a power series is constructed based on a fractional order basis function, which serves as the approximate solution. Secondly, neural networks and the initial conditions of differential equations are integrated into the construction of approximate solutions. Furthermore, the solution is discretized, and a multi-step unfolding strategy is employed on the resulting discrete solution. This approach ensures that each point in the solution is influenced by its predecessor. By means of repeated applications of the optimization algorithm, the residuals are successively diminished, thereby yielding approximate solutions to the equations. Finally, the efficacy and versatility of the proposed strategy were validated through a series of numerical experiments. Compared with the method of fractional physics-informed neural networks, there are up to <span><math><mn>18.7</mn></math></span>-fold and <span><math><mn>22.8</mn></math></span>-fold increases in stability of average and maximum residuals. Simultaneously, initial conditions are retained in new solutions.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"185 ","pages":"Article 113907"},"PeriodicalIF":6.6,"publicationDate":"2025-09-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145159665","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":"ORESTE methodology within a circular intuitionistic fuzzy framework for preferential outranking in hybrid cloud service selection","authors":"Ting-Yu Chen","doi":"10.1016/j.asoc.2025.113864","DOIUrl":"10.1016/j.asoc.2025.113864","url":null,"abstract":"<div><div>This paper advances the ORESTE (Organísation, Rangement Et Synthèse de Données Relarionnelles) methodology within the Circular Intuitionistic Fuzzy (CIF) framework, highlighting its potential in practical decision analytics. The study first enhances CIF aggregation by employing the generalized mean technique, offering a flexible way to combine evaluative ratings and significance weights. Through modulation of the averaging parameter, decision-makers are able to accentuate either lower or higher values, thereby overcoming the constraints associated with conventional arithmetic means. The framework further improves decision precision through CIF similarity-driven appraisal indices, which utilize refined similarity metrics grounded in axiomatic properties such as symmetry, boundedness, identity, and monotonicity. These indices quantify the similarity between evaluative ratings and anchor references, while also revealing indifference and incomparability—thus equipping decision-makers with a comprehensive toolset for handling uncertainty. The CIF ORESTE framework comprises two methodologies. CIF ORESTE I delivers a global weak ranking using similarity-driven indices and generalized projection-related distances. CIF ORESTE II addresses the limitations of weak rankings by incorporating an Indifference-Preference-Incomparability (I-P-R) structure, which uses mean and net preference intensities to establish thresholds and clarify outranking relations. Applied to the evaluation of hybrid cloud services for a technology corporation, the CIF ORESTE framework demonstrates its effectiveness in resolving group decisions, managing uncertainty, and structuring preferences. Comparative analyses further underscore its robustness in handling CIF-based data and delivering reliable results.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"185 ","pages":"Article 113864"},"PeriodicalIF":6.6,"publicationDate":"2025-09-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145183692","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 unsupervised framework for drift-aware anomaly detection in streaming time series","authors":"Danlei Li ,&nbsp;Nirmal-Kumar C. Nair,&nbsp;Kevin I-Kai Wang","doi":"10.1016/j.asoc.2025.113903","DOIUrl":"10.1016/j.asoc.2025.113903","url":null,"abstract":"<div><div>This paper presents an unsupervised adaptive drift-aware anomaly detection framework (ADA-ADF) designed to address the challenges of concept drift in time series data streams. ADA-ADF integrates a hybrid drift detection mechanism, combining statistical tests with performance-based metrics to accurately identify and distinguish between sudden and incremental drifts. To ensure effective adaptation, it employs a replay-based model update strategy that adjusts replay ratios in a drift-specific manner and incorporates representative historical data based on reconstruction errors. This approach allows the model to seamlessly adapt to evolving data distributions while maintaining high stability and accuracy. Extensive experiments on four diverse datasets demonstrate ADA-ADF’s superior performance in managing various drift and application scenarios. It consistently outperforms state-of-the-art methods, particularly in environments characterized by incremental or sudden drifts. With robust adaptability to changing data patterns and accurate anomaly detection capabilities, ADA-ADF provides a reliable solution for real-world applications, such as IoT and environmental monitoring.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"185 ","pages":"Article 113903"},"PeriodicalIF":6.6,"publicationDate":"2025-09-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145221558","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 MADM model using Frank operations based power aggregation operator under p,q-quasirung orthopair fuzzy sets for highway selection in war-plane landing","authors":"Sanjita Giri ,&nbsp;Sankar Kumar Roy ,&nbsp;Muhammet Deveci","doi":"10.1016/j.asoc.2025.113918","DOIUrl":"10.1016/j.asoc.2025.113918","url":null,"abstract":"<div><div>In military logistics and operational planning, selecting an optimal highway for war-plane landings and take-offs is a critical and strategic decision. This process involves several key factors that directly affect mission success, operational safety, and public security. Among the most important attributes are the highway’s straight and long stretch with sufficient width to accommodate war-plane landing distances, and its surface condition, which must be free from obstacles, debris, and damage. Low traffic density is crucial to avoid the risk of collisions during landing. Additionally, favourable weather conditions, proximity to military camps, availability of emergency services and fuel, and a secure and hazard-free surrounding terrain are essential for safe and efficient operations. These factors collectively form the backbone of a reliable and tactical approach to highway selection for military air operations. Thus, in order to assess and rank various options for the landing and take-off of war planes, a strong and trustworthy procedure for making decisions is required. The purpose of this experiment is to build a comprehensive structure in multi-attribute decision making environment, using suggested <span><math><mi>p</mi><mo>,</mo><mi>q</mi></math></span>-quasirung orthopair fuzzy Frank power averaging as well as <span><math><mi>p</mi><mo>,</mo><mi>q</mi></math></span>-quasirung orthopair fuzzy Frank power geometric operators to capture ambiguity and uncertainty in highway selection. Furthermore, <span><math><mi>p</mi><mo>,</mo><mi>q</mi></math></span>-quasirung orthopair fuzzy Frank power weighted aggregation along with <span><math><mi>p</mi><mo>,</mo><mi>q</mi></math></span>-quasirung orthopair fuzzy Frank power weighted geometric operators are implemented for integrating the distance as well as similarity measures. Finally, sensitivity analysis and a comparison with the present technique are included to further demonstrate the superiority and validity of the technique that is suggested.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"185 ","pages":"Article 113918"},"PeriodicalIF":6.6,"publicationDate":"2025-09-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145159675","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}