IEEE Transactions on Evolutionary Computation最新文献_第2页

Co-Evolutionary Dual-Tree Genetic Programming for Fine-Grained Flower Image Classification 基于协同进化双树遗传规划的细粒度花图像分类

IF 15.9 1区计算机科学

IEEE Transactions on Evolutionary Computation Pub Date : 2026-10-01 Epub Date: 2025-07-21 DOI: 10.1109/TEVC.2025.3590738

Qinyu Wang;Ying Bi;Bing Xue;Mengjie Zhang

{"title":"Co-Evolutionary Dual-Tree Genetic Programming for Fine-Grained Flower Image Classification","authors":"Qinyu Wang;Ying Bi;Bing Xue;Mengjie Zhang","doi":"10.1109/TEVC.2025.3590738","DOIUrl":"10.1109/TEVC.2025.3590738","url":null,"abstract":"fine-grained flower image classification (FGFIC) is challenging due to high similarities among species and variations within species, especially with limited training data. Existing genetic programming (GP) methods address this task using a single GP tree but show limited performance in flower localization. To overcome this limitation, this article introduces co-evolutionary dual tree genetic programming (CEDTGP) for FGFIC, employing a dual-tree structure that separates flower localization from feature learning for classification, thereby improving both localization and classification accuracy. A novel co-evolutionary strategy further optimizes the interaction between the dual trees, enhancing overall performance. Furthermore, a new distance measurement is proposed to refine the assessment of individuals with improved classification performance. Evaluations on various flower datasets with different numbers of classes demonstrate that CEDTGP outperforms popular GP and convolutional neural network (CNN) methods. Further analysis highlights the effectiveness of the distance measurement and the co-evolutionary strategy, and it also demonstrates the precision of the regions detected by the proposed approach.","PeriodicalId":13206,"journal":{"name":"IEEE Transactions on Evolutionary Computation","volume":"30 5","pages":"1848-1862"},"PeriodicalIF":15.9,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144684603","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}

引用次数: 0

Learn to Optimize Large-Scale Uncapacitated Facility Location Problems With an Adaptive Problem Reduction Strategy 学习用自适应问题缩减策略优化大规模无容量设施选址问题

IF 15.9 1区计算机科学

IEEE Transactions on Evolutionary Computation Pub Date : 2026-10-01 Epub Date: 2025-11-06 DOI: 10.1109/TEVC.2025.3630048

Hao Tong;Shuaixiang Zhang;Jialin Liu;Bo Yuan;Ke Tang;Xin Yao

{"title":"Learn to Optimize Large-Scale Uncapacitated Facility Location Problems With an Adaptive Problem Reduction Strategy","authors":"Hao Tong;Shuaixiang Zhang;Jialin Liu;Bo Yuan;Ke Tang;Xin Yao","doi":"10.1109/TEVC.2025.3630048","DOIUrl":"10.1109/TEVC.2025.3630048","url":null,"abstract":"Many large-scale combinatorial optimization problems (LSCOPs) are challenging to solve due to their high dimensionalities and complex search spaces. To deal with tens of thousands of decision variables, problem reduction approaches have proven effective in reducing the original problem instance to a more manageable size, thereby decreasing the dimensionality of problem instances before optimization. However, since the dimensionality of transformed problem instances remains fixed during the optimization process, the effectiveness of existing methods is highly dependent on the selected features. In this article, instead of only transforming the original problem instance once prior to the optimization, we propose a novel learning-based adaptive problem reduction (LAPR) framework to facilitate solving LSCOPs. Based on a learning model, the LAPR framework initially transforms the original problem instance into a low-dimensional one, and then dynamically increases the problem dimensionality when necessary during the optimization. To evaluate the effectiveness and efficiency of our framework, we have applied it to solve uncapacitated facility location problems (UFLPs) using three effective features to construct a machine learning model learning from small instances. Based on this model, the LAPR framework first identifies a subset of the most promising facilities to construct an initial reduced problem instance through a novel optimal-k estimation strategy. Then, it incrementally incorporates potentially promising facilities into candidate solutions during the evolutionary process to improve the optimization performance. Extensive experimental studies on a series of large-scale UFLP benchmarks demonstrate that our proposed LAPR framework significantly enhances the performance of existing solution methods and consistently outperforms state-of-the-art problem reduction frameworks.","PeriodicalId":13206,"journal":{"name":"IEEE Transactions on Evolutionary Computation","volume":"30 5","pages":"1952-1965"},"PeriodicalIF":15.9,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145454643","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}

引用次数: 0

Nanorobots-Assisted In Vivo Computational Methodology for Tumor Lesion Exploitation and Lesion Boundary Exploration 纳米机器人辅助肿瘤病灶开发和病灶边界探索的体内计算方法

IF 15.9 1区计算机科学

IEEE Transactions on Evolutionary Computation Pub Date : 2026-10-01 Epub Date: 2025-09-10 DOI: 10.1109/TEVC.2025.3608196

Shaolong Shi;Zeyuan Xu;Jiawei Zou;Weiqi Liu;Qiujun Huang

{"title":"Nanorobots-Assisted In Vivo Computational Methodology for Tumor Lesion Exploitation and Lesion Boundary Exploration","authors":"Shaolong Shi;Zeyuan Xu;Jiawei Zou;Weiqi Liu;Qiujun Huang","doi":"10.1109/TEVC.2025.3608196","DOIUrl":"10.1109/TEVC.2025.3608196","url":null,"abstract":"An innovative in vivo computational model that reframes the challenge of early tumor detection as an optimization problem has been developed recently. In this scope, the tumor lesion is viewed as the optimal solution that to be found by the computing agents (i.e., nanorobots). The tumor-triggered biological gradient field provides the information for the real-time path planning and fitness evaluation of nanorobots. The main emphasis of previous work has been on engineering tumor-targeting methodologies that address the challenges imposed by in vivo biological conditions. In this article, we investigate the issue of tumor lesion exploitation by controlling the motion of nanorobotic swarm, which aims to enhance the operational efficiency of nanorobots deployed for tumor lesion identification. Thus, an exploitation approach about the settings of nanorobots moving direction and step size is developed. Furthermore, inspired by the spontaneous motion property of nanorobots, we propose cooperation and competition mechanisms to overcome the challenge of lesion boundary exploration in this article. Some computational simulations are performed to validate the efficacy of this novel methodology.","PeriodicalId":13206,"journal":{"name":"IEEE Transactions on Evolutionary Computation","volume":"30 5","pages":"1791-1802"},"PeriodicalIF":15.9,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145035283","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}

引用次数: 0

Environmental Knowledge-Driven Transfer Strategy With Collaborative Learning for Dynamic Constrained Multiobjective Optimization 动态约束多目标优化环境知识驱动的协同学习迁移策略

IF 15.9 1区计算机科学

IEEE Transactions on Evolutionary Computation Pub Date : 2026-10-01 Epub Date: 2025-12-11 DOI: 10.1109/TEVC.2025.3643205

Zhi Liu;Wei Song;Carlos A. Coello Coello;Yaochu Jin;Shengxiang Yang;Shitong Wang

{"title":"Environmental Knowledge-Driven Transfer Strategy With Collaborative Learning for Dynamic Constrained Multiobjective Optimization","authors":"Zhi Liu;Wei Song;Carlos A. Coello Coello;Yaochu Jin;Shengxiang Yang;Shitong Wang","doi":"10.1109/TEVC.2025.3643205","DOIUrl":"10.1109/TEVC.2025.3643205","url":null,"abstract":"In recent years, transfer-learning-based dynamic constrained multiobjective optimization algorithms (TL-DCMOAs) have shown great promise in handling dynamic constrained multiobjective optimization problems (DCMOPs). In existing TL-DCMOAs, environmental similarity is typically assessed by measures based on Euclidean distance. However, such measures merely capture overall numerical differences between two environments, which are sensitive to local fluctuations in environmental information, potentially leading to misjudgments of environmental similarity. Besides, most existing TL-DCMOAs leverage historical constrained and unconstrained Pareto-optimal solutions (UPSs) as knowledge, but neglect relations and implicit solutions behind them for acquiring knowledge. To address the above issues, an environmental knowledge-driven transfer strategy with collaborative learning (EKTS-CL) is proposed in this article. Specifically, we design two structural similarity measures (SSMs) which consider the mean, variance, and covariance of environmental information, to comprehensively assess the environmental structural discrepancies with respect to concentration, divergence, and correlation, respectively. Using SSMs, the most structurally similar historical UPSs and constrained Pareto-optimal solutions (CPSs) can be obtained. Additionally, true CPS and UPS share partial solutions across all environments or the majority of environments of a DCMOP. From the obtained CPS and UPS, that separately approximate the true CPS and UPS of the current environment, we extract implicit shared solutions and exclusive ones through collaborative learning, providing high-quality knowledge than the obtained plain solutions for knowledge transfer. Our experimental results include comprehensive comparisons with seven state-of-the-art DCMOAs, and show that the proposed EKTS-CL brings significant performance improvements in solving DCMOPs.","PeriodicalId":13206,"journal":{"name":"IEEE Transactions on Evolutionary Computation","volume":"30 5","pages":"2152-2167"},"PeriodicalIF":15.9,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145728831","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}

引用次数: 0

Space Net Optimization for Single Objective Bound Constrained Problems 单目标有界约束问题的空间网优化

IF 15.9 1区计算机科学

IEEE Transactions on Evolutionary Computation Pub Date : 2026-10-01 Epub Date: 2025-08-15 DOI: 10.1109/TEVC.2025.3599438

Chun-Wei Tsai;Yi-Cheng Yang;Tzu-Chieh Tang;Che-Wei Hsu

{"title":"Space Net Optimization for Single Objective Bound Constrained Problems","authors":"Chun-Wei Tsai;Yi-Cheng Yang;Tzu-Chieh Tang;Che-Wei Hsu","doi":"10.1109/TEVC.2025.3599438","DOIUrl":"https://doi.org/10.1109/TEVC.2025.3599438","url":null,"abstract":"Metaheuristic algorithms are capable of solving various complex engineering optimization problems efficiently; however, most of them rely on a certain number of searched solutions to guide their searches during the convergence process. Although it is not hard to imagine that a larger number of searched solutions will provide a better description of the landscape of the solution space, retaining all the searched solutions remains a challenge. Inspired by this observation, we present in this article a novel metaheuristic algorithm called space net optimization (SNO), which is equipped with a new mechanism called space net for depicting the landscape of the solution space using the information provided by most searched solutions. With the space net, a metaheuristic algorithm is then kind of like having a “vision” for knowing the landscape of the solution space to improve its search accuracy, such as investing more searches on high potential regions and reducing the number of searches on low potential regions. Simulation results show that SNO is capable of providing a promising result for solving single objective bound constrained problems.","PeriodicalId":13206,"journal":{"name":"IEEE Transactions on Evolutionary Computation","volume":"30 5","pages":"1833-1847"},"PeriodicalIF":15.9,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11126067","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149097539","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A Fitness Spatially Informed Evolutionary Algorithm for Deceptive Multimodal Multiobjective Optimization 欺骗性多模态多目标优化的适应度空间进化算法

IF 15.9 1区计算机科学

IEEE Transactions on Evolutionary Computation Pub Date : 2026-10-01 Epub Date: 2025-10-07 DOI: 10.1109/TEVC.2025.3618608

Yaqi Ti;Changhe Li;Xinye Cai;Shengxiang Yang

{"title":"A Fitness Spatially Informed Evolutionary Algorithm for Deceptive Multimodal Multiobjective Optimization","authors":"Yaqi Ti;Changhe Li;Xinye Cai;Shengxiang Yang","doi":"10.1109/TEVC.2025.3618608","DOIUrl":"10.1109/TEVC.2025.3618608","url":null,"abstract":"evolutionary algorithms (EAs) are effective for solving multimodal multiobjective optimization problem which optimal solutions subsets distributed regularly in the decision space. Existing EAs, however, are hard to achieve good diversity in the decision space due to their lack of thinking-in-space when solving deceptive multimodal multiobjective problems, and one manifestation of the deceptiveness is that global optima are composed of unconnected small optimal solution subsets that randomly distributed in the decision space with many local optima subsets mixed among. In this study, for enabling EAs to continuously search without trapping, a solving pathway is proposed that identifies information-rich historical data and constructs structured spatial representation based on high-quality evolutionary data. Furthermore, a moderate region-oriented solution set update methodology is proposed for the detection of ultrasmall optimal modalities, and two novel environmental selection and archive update mechanisms are designed. A Memory-aiding Search multimodal multiobjective Optimization EA, i.e., MS-multimodal multiobjective optimization EA (MMOEA), is instantiated. The experimental results on 27 deceptive instances and a real-world problem demonstrate MS-MMOEA’s competitiveness over the state-of-the-art.","PeriodicalId":13206,"journal":{"name":"IEEE Transactions on Evolutionary Computation","volume":"30 5","pages":"1923-1936"},"PeriodicalIF":15.9,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145241480","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}

引用次数: 0

Fully Tensorized GPU-Accelerated Multipopulation Evolutionary Algorithm for Constrained Multiobjective Optimization Problems 约束多目标优化问题的全张化gpu加速多种群进化算法

IF 15.9 1区计算机科学

IEEE Transactions on Evolutionary Computation Pub Date : 2026-10-01 Epub Date: 2026-01-12 DOI: 10.1109/TEVC.2026.3651395

Weixiong Huang;Rui Wang;Wenhua Li;Sheng Qi;Tianyu Luo;Delong Chen;Tao Zhang;Ling Wang

{"title":"Fully Tensorized GPU-Accelerated Multipopulation Evolutionary Algorithm for Constrained Multiobjective Optimization Problems","authors":"Weixiong Huang;Rui Wang;Wenhua Li;Sheng Qi;Tianyu Luo;Delong Chen;Tao Zhang;Ling Wang","doi":"10.1109/TEVC.2026.3651395","DOIUrl":"10.1109/TEVC.2026.3651395","url":null,"abstract":"Real-world constrained multiobjective optimization problems (CMOPs) are prevalent and often come with stringent time-sensitive requirements. However, most contemporary constrained multiobjective evolutionary algorithms (CMOEAs) suffer from a number of drawbacks, including complex designs, low computational efficiency, and long convergence times, which are particularly pronounced when addressing time-sensitive CMOPs. Although research on accelerating evolutionary algorithms (EAs) using GPU parallelism has advanced, existing CMOEAs still face significant limitations within GPU frameworks. To overcome these challenges, this article proposes a GPU-accelerated multipopulation EA, termed GMPEA. We first systematically analyze the performance bottlenecks of representative CMOEAs when implemented in a GPU environment. To address the tradeoff between computational speed and solution performance, GMPEA introduces a decomposition-based multipopulation approach that is fully parallelized across its entire workflow. We conducted comparative experiments on various benchmark tests and real-world applications—the Weapon Target Assignment Problems. The results demonstrate that GMPEA achieves competitive performance even without time constraints, while its computational speed significantly surpasses that of the compared algorithms. More critically, under a strict time limit, GMPEA’s performance drastically outperforms its counterparts. This work provides compelling evidence of GMPEA’s superiority in solving time-sensitive CMOPs. The source code is available at <uri>https://github.com/785509490/EvoCmo</uri>.","PeriodicalId":13206,"journal":{"name":"IEEE Transactions on Evolutionary Computation","volume":"30 5","pages":"2185-2199"},"PeriodicalIF":15.9,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145955237","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}

引用次数: 0

A Multimodal Ant Colony Optimization Framework for Multisolution Traveling Salesman Problem 多解旅行商问题的多模态蚁群优化框架

IF 15.9 1区计算机科学

IEEE Transactions on Evolutionary Computation Pub Date : 2026-10-01 Epub Date: 2025-11-28 DOI: 10.1109/TEVC.2025.3638360

Han-Yang Pan;Ming Li;Qiang Yang;Wei-Neng Chen;Ya-Hui Jia;Jian-Yu Li;Dan-Ting Duan;Dong Liu;Jun Zhang

{"title":"A Multimodal Ant Colony Optimization Framework for Multisolution Traveling Salesman Problem","authors":"Han-Yang Pan;Ming Li;Qiang Yang;Wei-Neng Chen;Ya-Hui Jia;Jian-Yu Li;Dan-Ting Duan;Dong Liu;Jun Zhang","doi":"10.1109/TEVC.2025.3638360","DOIUrl":"10.1109/TEVC.2025.3638360","url":null,"abstract":"This article devises a multimodal ACO (MACO) framework to solve multisolution traveling salesman problems (MSTSPs), which aim to find multiple optimal routes in a single run. Specifically, a dynamic niching strategy is designed to partition the ant colony into exclusive niches based on the route similarity measured by the common edges between two routes. Subsequently, MACO maintains a pheromone matrix for each niche and lets the ants in each niche construct their routes independently based on the associated pheromone matrix. For effective pheromone update, this article designs a niche and pheromone matrix association rule to match each newly partitioned niche with one pheromone matrix. To further improve optimization efficiency, this article designs a Dynamic Critical Edge Strategy to dig out critical edges during the optimization. Then, these critical edges are cooperated with 2-opt to refine the route constructed by each ant. To preserve the found optimal routes, this article devises an optimal route preservation strategy to store as many different high-quality routes as possible. Lastly, we instantiate MACO with four canonical ACO algorithms, namely Ant System(AS), Elite AS(EAS), Rank-based AS(ASRank), and Max-Min AS(MMAS), and then conduct extensive experiments to compare them with six state-of-the-art algorithms on the only one MSTSP benchmark suite and the TSPLIB instances. The experimental results reveal that the four MACOs perform competitively with one latest algorithm, but significantly outperform the other five comparison algorithms. Further analysis discovers that ASRank is the most effective one among the four ant colony optimizations (ACOs) to cooperate with MACO in solving MSTSPs.","PeriodicalId":13206,"journal":{"name":"IEEE Transactions on Evolutionary Computation","volume":"30 5","pages":"2168-2184"},"PeriodicalIF":15.9,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145613448","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}

引用次数: 0

Fuzzy Logic-Based Multifactorial Optimization for Evolutionary Multitasking 基于模糊逻辑的多因素进化多任务优化

IF 15.9 1区计算机科学

IEEE Transactions on Evolutionary Computation Pub Date : 2026-10-01 Epub Date: 2025-12-09 DOI: 10.1109/TEVC.2025.3641950

Wu Lin;Songbai Liu;Qiuzhen Lin;Liang Feng;Kay Chen Tan

{"title":"Fuzzy Logic-Based Multifactorial Optimization for Evolutionary Multitasking","authors":"Wu Lin;Songbai Liu;Qiuzhen Lin;Liang Feng;Kay Chen Tan","doi":"10.1109/TEVC.2025.3641950","DOIUrl":"10.1109/TEVC.2025.3641950","url":null,"abstract":"Multifactorial optimization (MFO) has gained increasing attention in evolutionary multitasking (EMT), as it enables implicit knowledge transfer among tasks to enhance overall optimization efficiency. However, selecting appropriate transfer methods and adjusting their transfer probabilities remain significant challenges in MFO. Since existing selection and adjustment rules are typically formulated based on classical logic, they often fail to effectively handle the inherent uncertainty and dynamic nature of EMT. To address these limitations, this article suggests a new multifactorial evolutionary algorithm based on fuzzy logic, called MFEA-FL. By employing the Mamdani fuzzy inference system (FIS), expert knowledge can be naturally incorporated into the decision-making process, enabling MFEA-FL to exhibit strong fault tolerance and robustness across a wide range of multitasking environments. Specifically, two strategies, namely fuzzy transfer method selection (FTMS) and fuzzy transfer probability adjustment (FTPA), are proposed. FTMS quantifies the applicability of candidate transfer methods via a Mamdani FIS and then adaptively selects the most promising one for conducting knowledge transfer. Meanwhile, FTPA estimates the change of the transfer probability via a Mamdani FIS and dynamically adjusts the intensity of knowledge transfer. Experimental results on a series of multitasking optimization problems validate the effectiveness of the two fuzzy strategies and the superior performance of MFEA-FL over existing state-of-the-art EMT algorithms.","PeriodicalId":13206,"journal":{"name":"IEEE Transactions on Evolutionary Computation","volume":"30 5","pages":"2108-2122"},"PeriodicalIF":15.9,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145717956","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}

引用次数: 0

Collaborative Evolutionary Multitasking Multiobjective Adversarial Attacks in Hyperspectral Image Classification 高光谱图像分类中的协同进化多任务多目标对抗攻击

IF 15.9 1区计算机科学

IEEE Transactions on Evolutionary Computation Pub Date : 2026-10-01 Epub Date: 2025-08-18 DOI: 10.1109/TEVC.2025.3599560

Hao Li;Jingying Li;Maoguo Gong;Kelin Dang;A.K. Qin;Yue Wu;Lining Xing

{"title":"Collaborative Evolutionary Multitasking Multiobjective Adversarial Attacks in Hyperspectral Image Classification","authors":"Hao Li;Jingying Li;Maoguo Gong;Kelin Dang;A.K. Qin;Yue Wu;Lining Xing","doi":"10.1109/TEVC.2025.3599560","DOIUrl":"https://doi.org/10.1109/TEVC.2025.3599560","url":null,"abstract":"Adversarial attacks on hyperspectral image (HSI) classification has garnered significant attention. However, most of the existing adversarial attack methods tend to apply global perturbations to the HSI, ignoring the fact that different classes in the HSI have distinct preferences for the spectral bands, and fails to generate an approximate optimal perturbation for each class. Therefore, this article establishes an evolutionary multitasking multiobjective adversarial attack model to generate a universal adversarial perturbation for each class. Then, we propose a novel collaborative evolutionary multitasking optimization (CoEMTO) method to optimize all tasks collaboratively. In CoEMTO, the optimal solution is the combination of solutions for all tasks, and a dynamic reward resource allocation strategy is designed to approximate the optimal solution for all tasks. In addition, perturbations are applied to the abundance rather than directly manipulating the pixel spectra to ensure the physical consistency of the generated adversarial perturbations. Experimental results demonstrate that the proposed CoEMTO algorithm can efficiently converge toward the Pareto front, and significantly outperforms other adversarial attack methods in terms of attack effectiveness.","PeriodicalId":13206,"journal":{"name":"IEEE Transactions on Evolutionary Computation","volume":"30 5","pages":"1863-1877"},"PeriodicalIF":15.9,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149097945","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}

引用次数: 0