Swarm and Evolutionary Computation最新文献

Bi-objective evolutionary hyper-heuristics in automated machine learning for text classification tasks 文本分类任务自动机器学习中的双目标进化超启发式

IF 8.2 1区计算机科学

Swarm and Evolutionary Computation Pub Date : 2025-07-25 DOI: 10.1016/j.swevo.2025.102073

Jonathan Estrella-Ramírez , Jorge de la Calleja , Juan Carlos Gómez Carranza

{"title":"Bi-objective evolutionary hyper-heuristics in automated machine learning for text classification tasks","authors":"Jonathan Estrella-Ramírez , Jorge de la Calleja , Juan Carlos Gómez Carranza","doi":"10.1016/j.swevo.2025.102073","DOIUrl":"10.1016/j.swevo.2025.102073","url":null,"abstract":"<div><div>This paper proposes an evolutionary model based on hyper-heuristics to automate the selection of classification methods for text datasets under a bi-objective approach. The model has three nested levels. At the first level, individual methods classify datasets, recording two performances: the number of misclassifications and computational time, which are often in conflict. At the second level, hyper-heuristics, as a set of rules of the form <span><math><mrow><mi>i</mi><mi>f</mi><mo>→</mo><mi>t</mi><mi>h</mi><mi>e</mi><mi>n</mi></mrow></math></span>, select classification methods for datasets based on 16 meta-features representing the data distribution. The fitness for a hyper-heuristic is evaluated on a training group of datasets by aggregating the two low-level performances of the chosen methods. At the third level, the multi-objective evolutionary algorithm Strength Pareto Evolutionary Algorithm 2 evolves hyper-heuristic populations considering the bi-objective of minimizing the two aggregated performances. The result is a Pareto-approximated front of hyper-heuristics, which offers a range of solutions from computationally efficient to high classification performance. Finally, the model evaluates the front with an independent test group of datasets and selects those hyper-heuristics that are not dominated. We evaluated the resulting fronts through extensive experiments, measuring several quality indicators. We compare the model’s fronts with a front baseline consisting of non-dominated individual classification methods and four state-of-the-art automated machine learning tools (AutoKeras, AutoGluon, H2O, and TPOT). The proposed model yields larger, more diverse Pareto-approximated fronts that outperform the baseline front, allowing solution selection based on available resources and trade-offs between performance and cost.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"98 ","pages":"Article 102073"},"PeriodicalIF":8.2,"publicationDate":"2025-07-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144702703","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 non-index mixed-asset portfolio optimization approach 一种非指数混合资产组合优化方法

IF 8.2 1区计算机科学

Swarm and Evolutionary Computation Pub Date : 2025-07-25 DOI: 10.1016/j.swevo.2025.102074

Yuyang Bai , Changsheng Zhang , Baiqing Sun , Bin Zhang

{"title":"A non-index mixed-asset portfolio optimization approach","authors":"Yuyang Bai , Changsheng Zhang , Baiqing Sun , Bin Zhang","doi":"10.1016/j.swevo.2025.102074","DOIUrl":"10.1016/j.swevo.2025.102074","url":null,"abstract":"<div><div>As the financial industry shifts from divided operations to mixed operations, mixed-asset portfolios have gradually gained ground in investment portfolios. Existing mixed-asset portfolio optimization approaches frequently introduce indices for representing asset classes to eliminate heterogeneity among asset classes. However, few introduced indices comprehensively and realistically represent asset classes, leading to a loss of feasible solutions and practical reliability. To address these limitations, this paper proposes a non-index mixed-asset portfolio optimization approach consisting of problem modeling and problem solving. For problem modeling, our approach models the mixed-asset portfolio optimization as a multi-objective bi-level optimization problem. In the inner-level optimization, optimal portfolios within each asset class are constructed to represent the corresponding asset class. These optimal portfolios contain more information and are constructed from realistically available products, thus representing the asset class more comprehensively and practically. In the outer-level optimization, the allocation among the asset classes is optimized to obtain an optimal mixed-asset portfolio. For problem solving, a multi-swarm dynamic cooperative optimization method is proposed to solve the modeled problem. Considering that obtaining the complete inner-level optimization of the modeled problem is challenging and time-consuming, a dynamic collaboration mechanism is designed to obtain the optimal subset of the inner-level optimization, thus solving the problem efficiently and effectively. To verify the effectiveness of our proposed non-index approach, an experiment is conducted to compare our proposed approach with four state-of-the-art approaches. Our proposed non-index approach problem outperforms competitors in 27 of 30 scenarios on both the Pareto optimality and the realistic performance.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"98 ","pages":"Article 102074"},"PeriodicalIF":8.2,"publicationDate":"2025-07-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144702689","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 frequent pattern-based coevolutionary framework for multi-component spectral feature selection 基于频繁模式的多组分光谱特征选择协同进化框架

IF 8.2 1区计算机科学

Swarm and Evolutionary Computation Pub Date : 2025-07-24 DOI: 10.1016/j.swevo.2025.102077

Panpan Zhang , Jing Rong , Ye Tian , Yajie Zhang , Shangshang Yang , Xingyi Zhang

{"title":"A frequent pattern-based coevolutionary framework for multi-component spectral feature selection","authors":"Panpan Zhang , Jing Rong , Ye Tian , Yajie Zhang , Shangshang Yang , Xingyi Zhang","doi":"10.1016/j.swevo.2025.102077","DOIUrl":"10.1016/j.swevo.2025.102077","url":null,"abstract":"<div><div>Spectral feature selection plays a crucial role in spectral analysis as it aims to identify the most effective features from the original high-dimensional wavelength variables, thereby enhancing the accuracy of concentration prediction models. In multi-component spectral feature selection (MCSFS) problems, diverse composition and concentration of samples result in complex overlapping peaks and correlations among variables. This complexity poses challenges in finding optimal subsets of features efficiently. To address this issue, this paper proposes a frequent pattern-based coevolutionary framework for solving MCSFS problems. Specifically, the algorithm starts by generating a main population for multi-component spectral feature selection and multiple auxiliary populations for single-component spectral feature selection. Furthermore, it introduces a frequent pattern mining strategy to identify dynamic superior feature combinations and their updated weights in each population, dealing with the complexity of variables to accelerate the search for effective features. The proposed coevolutionary framework facilitates interactions between populations by sharing the identified feature combinations and offspring information, leading to the acquisition of high-quality feature selection results. Experimental results on twelve MCSFS problems, based on three high-dimensional spectral datasets, demonstrate that the proposed algorithm outperforms six state-of-the-art evolutionary algorithms.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"98 ","pages":"Article 102077"},"PeriodicalIF":8.2,"publicationDate":"2025-07-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144702682","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

Distributed assembly flexible job shop scheduling with dual-resource constraints via a deep Q-network based memetic algorithm 基于深度q网络模因算法的双资源约束分布式装配柔性作业车间调度

IF 8.2 1区计算机科学

Swarm and Evolutionary Computation Pub Date : 2025-07-24 DOI: 10.1016/j.swevo.2025.102086

Hongliang Zhang , Yi Chen , Gongjie Xu , Yuteng Zhang

{"title":"Distributed assembly flexible job shop scheduling with dual-resource constraints via a deep Q-network based memetic algorithm","authors":"Hongliang Zhang , Yi Chen , Gongjie Xu , Yuteng Zhang","doi":"10.1016/j.swevo.2025.102086","DOIUrl":"10.1016/j.swevo.2025.102086","url":null,"abstract":"<div><div>The distributed flexible job shop scheduling problem (DFJSP) has garnered significant attention due to the shift of production paradigms. However, existing research of DFJSP primarily focuses on machine resources while neglecting worker resources, which play a crucial role in enhancing productivity. Furthermore, manufacturing processes often involve both processing and assembly stages. An integrated approach to scheduling the two stages can significantly enhance efficiency and reduce costs. This study addresses the distributed assembly flexible job shop scheduling problem with dual resource constraints (DAFJSP-DRC), aiming to minimize total product tardiness (TPT), total energy consumption (TEC), and total cost (TCO). To tackle this problem, we develop a mixed-integer programming (MIP) model and propose a deep Q-network-based memetic algorithm (DQNMA). In DQNMA, high-quality initial solutions are generated based on processing resources, and a two-stage decoding mechanism is designed to get efficient scheduling schemes. Then, crossover and mutation operators for critical factories are proposed, and a deep Q-network is designed to dynamically adjust the crossover and mutation rates. Furthermore, eight neighborhood structures are designed to enhance solution diversity, while a tabu search-based local search strategy improves the algorithm's exploration and exploitation capabilities. Eventually, extensive experimental results demonstrate the effectiveness of the proposed strategies in enhancing the performance of DQNMA. Comparative analysis against four state-of-the-art multi-objective algorithms validates the superiority and effectiveness of the designed algorithm.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"98 ","pages":"Article 102086"},"PeriodicalIF":8.2,"publicationDate":"2025-07-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144702683","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 self-adjusting representation-based multitask PSO for high-dimensional feature selection 基于自调整表示的多任务粒子群高维特征选择

IF 8.2 1区计算机科学

Swarm and Evolutionary Computation Pub Date : 2025-07-23 DOI: 10.1016/j.swevo.2025.102084

Li Deng , Xiaohui Su , Bo Wei

{"title":"A self-adjusting representation-based multitask PSO for high-dimensional feature selection","authors":"Li Deng , Xiaohui Su , Bo Wei","doi":"10.1016/j.swevo.2025.102084","DOIUrl":"10.1016/j.swevo.2025.102084","url":null,"abstract":"<div><div>As a critical preprocessing step in machine learning tasks, feature selection (FS) aims to identify informative features from the original datasets. However, FS is commonly formulated as an NP-hard combinatorial optimization problem, particularly when compounded by exponentially expanding search spaces and complex feature interactions. Due to its simplicity of implementation, Particle Swarm Optimization (PSO) has been extensively utilized in FS tasks. Concurrently, the optimization process frequently converges to local optima when handling high-dimensional (<span><math><mo>></mo></math></span>1000D) datasets. To address this issue, a self-adjusting representation-based multitask PSO (SAR-MTPSO) is proposed in this paper. Firstly, the knee point strategy and an elite feature-preserving strategy are employed to obtain promising particles with key features. Based on these particles, a new multitask framework is introduced, where two tasks are constructed by using a self-adjusting representation strategy. Secondly, a two-layer knowledge transfer strategy is proposed to promote the useful information sharing and exchanging between two tasks dynamically. Finally, an adaptive re-initialization strategy is proposed to enhance the exploitation and exploration capabilities of the two tasks respectively. SAR-MTPSO was compared with 10 representative FS algorithms on 21 high-dimensional datasets. Experimental results show that SAR-MTPSO can achieve the highest classification accuracies with smaller sizes of feature subsets on 17 out of 21 datasets.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"98 ","pages":"Article 102084"},"PeriodicalIF":8.2,"publicationDate":"2025-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144686483","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

Experience Exchange Strategy: An evolutionary strategy for meta-heuristic optimization algorithms 经验交换策略：元启发式优化算法的进化策略

IF 8.2 1区计算机科学

Swarm and Evolutionary Computation Pub Date : 2025-07-23 DOI: 10.1016/j.swevo.2025.102082

Heming Jia , Honghua Rao

{"title":"Experience Exchange Strategy: An evolutionary strategy for meta-heuristic optimization algorithms","authors":"Heming Jia , Honghua Rao","doi":"10.1016/j.swevo.2025.102082","DOIUrl":"10.1016/j.swevo.2025.102082","url":null,"abstract":"<div><div>Meta-heuristic optimization algorithms typically change individual positions based on iterations, causing the population to switch search regions. This may result in the original search area not being explored in depth, thereby reducing the optimization performance of the algorithm. To deepen the connection between populations and individuals, this article proposes an evolutionary strategy called Experience Exchange Strategy (EES). EES considers the relationship between individuals and populations, deepening the connection between individuals and populations. EES has structured into three distinct stages: the experience scarcity stage (ESC), the experience crossover stage (ECR), the experience sharing stage (ESH). In the ESC, due to many areas not being searched, the population lacks search experience and mainly relies on primitive algorithms to find positions. This can preserve the optimization effect of the original algorithm and explore more positions. In the ECR, due to the accumulation of more experience in the population, individuals will update their positions based on more reference population experience. This can improve the accuracy of the search range and conduct more detailed searches. In the ESH, the population accumulates a large amount of experience, and individuals conduct more detailed searches based on the population’s experience. Through ESH, the population can search intensively to find a better position more finely. To verify the performance of EES, this article conducted optimization tests using IEEE CEC2014 and IEEE CEC2020 functions. And 15 algorithms were selected for improvement and compared with the original algorithm. Then, 57 single objective constrained engineering problems were used for testing experiments. The experimental results demonstrate that EES significantly improves the performance of meta-heuristic optimization algorithms.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"98 ","pages":"Article 102082"},"PeriodicalIF":8.2,"publicationDate":"2025-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144686484","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 constrained multi-objective coevolutionary algorithm with adaptive operator selection for efficient test scheduling in interposer-based 2.5D ICs 基于中介体的2.5D集成电路测试调度的约束多目标自适应协同进化算法

IF 8.2 1区计算机科学

Swarm and Evolutionary Computation Pub Date : 2025-07-23 DOI: 10.1016/j.swevo.2025.102085

Chunlei Li , Libao Deng , Liyan Qiao , Lili Zhang

{"title":"A constrained multi-objective coevolutionary algorithm with adaptive operator selection for efficient test scheduling in interposer-based 2.5D ICs","authors":"Chunlei Li , Libao Deng , Liyan Qiao , Lili Zhang","doi":"10.1016/j.swevo.2025.102085","DOIUrl":"10.1016/j.swevo.2025.102085","url":null,"abstract":"<div><div>Interposer-based 2.5-dimensional integrated circuits (2.5D ICs) have emerged as a promising solution to address wire delay and power consumption challenges in modern semiconductor design. However, the increasing complexity and density of 2.5D ICs introduces critical test scheduling challenges, where existing methods fail to effectively optimize hardware cost and test time while satisfying strict power and duration constraints. To overcome these limitations, this paper models the test scheduling problem in 2.5D ICs as a constrained multi-objective optimization problem (CMOP) and proposes a constrained multi-objective coevolutionary algorithm (termed AOSCEA) with adaptive operator selection. The algorithm introduces a two-chromosome-based encoding method paired with a matching-level-based decoding strategy to effectively map the discrete scheduling problem to continuous evolutionary algorithm frameworks, enabling efficient exploration of the search space. A coevolutionary mechanism is incorporated into the algorithm with two populations: a main population that solves the CMOP and an auxiliary population that ignores constraints to enhance exploration. Additionally, targeting to enhance the versatility of the algorithm across different test scheduling problems, AOSCEA employs two deep <em>Q</em>-networks to adaptively select genetic operators and constraint handling techniques for the main population during the optimization process. Extensive experiments on various test scheduling instances in 2.5D ICs with different scales demonstrate that AOSCEA outperforms several state-of-the-art algorithms in terms of solution quality, convergence speed, and robustness.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"98 ","pages":"Article 102085"},"PeriodicalIF":8.2,"publicationDate":"2025-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144686482","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

Learning strategies for particle swarm optimizer: A critical review and performance analysis 粒子群优化器的学习策略：一个关键的回顾和性能分析

IF 8.2 1区计算机科学

Swarm and Evolutionary Computation Pub Date : 2025-07-22 DOI: 10.1016/j.swevo.2025.102048

Dikshit Chauhan , Shivani , Ponnuthurai N. Suganthan

{"title":"Learning strategies for particle swarm optimizer: A critical review and performance analysis","authors":"Dikshit Chauhan , Shivani , Ponnuthurai N. Suganthan","doi":"10.1016/j.swevo.2025.102048","DOIUrl":"10.1016/j.swevo.2025.102048","url":null,"abstract":"<div><div>Nature has long inspired the development of swarm intelligence (SI), a key branch of artificial intelligence that models collective behaviors observed in biological systems for solving complex optimization problems. Particle swarm optimization (PSO) is widely adopted among SI algorithms due to its simplicity and efficiency. Despite numerous learning strategies proposed to enhance PSO’s performance in terms of convergence speed, robustness, and adaptability, no comprehensive and systematic analysis of these strategies exists. We review and classify various learning strategies to address this gap, assessing their impact on optimization performance. Additionally, a comparative experimental evaluation is conducted to examine how these strategies influence PSO’s search dynamics. Our analysis reveals that multi-swarm strategies consistently outperform other PSO strategies in high-dimensional and multimodal problems, offering better exploration and convergence trade-offs. Finally, we discuss open challenges and future directions, emphasizing the need for self-adaptive, intelligent PSO variants capable of addressing increasingly complex real-world problems. This survey not only synthesizes the current landscape of learning-enhanced PSO but also provides actionable insights for future research and algorithmic design.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"98 ","pages":"Article 102048"},"PeriodicalIF":8.2,"publicationDate":"2025-07-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144679757","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

Weak-prior medical image matting based on microscale-searching evolutionary optimization 基于微尺度搜索进化优化的弱先验医学图像抠图

IF 8.2 1区计算机科学

Swarm and Evolutionary Computation Pub Date : 2025-07-22 DOI: 10.1016/j.swevo.2025.102065

Li Kang , Han Huang , Yihui Liang , Baoxiong Zhuang

{"title":"Weak-prior medical image matting based on microscale-searching evolutionary optimization","authors":"Li Kang , Han Huang , Yihui Liang , Baoxiong Zhuang","doi":"10.1016/j.swevo.2025.102065","DOIUrl":"10.1016/j.swevo.2025.102065","url":null,"abstract":"<div><div>Image matting is to predict alpha mattes that reflects the opacity of images, recently showing potential in identifying transition regions of lesions in computer aided diagnosis. Image matting can be modeled as a large-scale combinatorial optimization problem that has numerous subproblems. Evolutionary algorithms (EAs) have been applied to predict accurate alpha mattes. The advantage of EAs-based methods is the ability to predict alpha mattes with weak prior like trimaps that provide value of opacity for pixels effortless to annotate compared to recent deep learning-based methods. However, it is challenging for EAs to solve the problem efficiently due to numerous subproblems and the large size of the decision set. Based on the observation that the similarity of subproblems correlates with the similarity of their objective spaces, this paper proposes a method for estimating a microscale subset of the decision set from the solving process of similar subproblems. A framework is designed to reduce the exploration cost of EAs in the large-scale decision set by guiding EAs to search in this estimated microscale subsets. Three medical image matting datasets are used to validate our method’s improvement in the efficiency of evolutionary algorithms. Experimental results demonstrate that EAs embedded in the proposed framework obtain the best prediction of alpha mattes on medical images and also in weak scenarios involving natural images. Comparative experimental results on multi-objective performance metrics indicate that our method is capable of finding superior solutions using fewer fitness evaluations. The contribution of our work is to make EAs an efficient approach to solving the medical image matting problem with weak prior.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"98 ","pages":"Article 102065"},"PeriodicalIF":8.2,"publicationDate":"2025-07-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144679758","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

Dynamic multi-objective optimization algorithm based on incremental Gaussian mixture model 基于增量高斯混合模型的动态多目标优化算法

IF 8.2 1区计算机科学

Swarm and Evolutionary Computation Pub Date : 2025-07-21 DOI: 10.1016/j.swevo.2025.102067

Xuewen Xia, Yi Zeng, Xing Xu, Yinglong Zhang

{"title":"Dynamic multi-objective optimization algorithm based on incremental Gaussian mixture model","authors":"Xuewen Xia, Yi Zeng, Xing Xu, Yinglong Zhang","doi":"10.1016/j.swevo.2025.102067","DOIUrl":"10.1016/j.swevo.2025.102067","url":null,"abstract":"<div><div>In recent years, prediction-based algorithms have made significant progress in solving dynamic multi-objective optimization problems (DMOPs). However, most existing methods only consider information from several consecutive environments and ignore past search experiences. To address the issue, this paper proposes a novel dynamic multi-objective evolutionary algorithm (DMOEA) based on an incremental Gaussian mixture model (IGMM). When environmental changes occurred, the initial population in the new environment is composed of two parts. The one part includes a few predicted individuals generated by IGMM aiming to explore the potential correlation between environments. To ensure quality of the individuals generated by IGMM, a feature-based augmentation strategy is employed to generate representative training data before the training process of IGMM. The other part consists of some individuals created via polynomial mutation operator based on randomly selected solutions from the previous environment. Based on the hybrid initial population, IGMM-DMOEA can quickly respond to environmental changes. To testify the performance of IGMM-DMOEA, twenty widely used benchmark functions and three real-world applications are adopted in this study. Extensive experimental results verify that IGMM-DMOEA can exhibit effective response to environmental changes. Comparisons results between it and other seven peer algorithms suggest that IGMM-DMOEA attains more reliable performance, measured by three popular metrics. Moreover, the effectiveness and efficiency of the new proposed strategies are discussed based on extensive experiments.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"98 ","pages":"Article 102067"},"PeriodicalIF":8.2,"publicationDate":"2025-07-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144672119","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