Swarm and Evolutionary Computation最新文献

{"title":"An individual adaptive evolution and regional collaboration based evolutionary algorithm for large-scale constrained multiobjective optimization problems","authors":"Kunjie Yu ,&nbsp;Zhenyu Yang ,&nbsp;Jing Liang ,&nbsp;Kangjia Qiao ,&nbsp;Boyang Qu ,&nbsp;Ponnuthurai Nagaratnam Suganthan","doi":"10.1016/j.swevo.2025.101925","DOIUrl":"10.1016/j.swevo.2025.101925","url":null,"abstract":"<div><div>Large-scale constrained multiobjective optimization problems (LSCMOPs) refer to constrained multiobjective optimization problems (CMOPs) with large-scale decision variables. When using evolutionary algorithms to solve LSCMOPs, the main challenge lies in balancing feasibility, convergence, and diversity in the high-dimensional search space. However, only a few studies focus on LSCMOPs and most existing related algorithms fail to achieve satisfactory performance. This paper proposes two novel mechanisms (the individual adaptive evolution strategy and the regional collaboration mechanism) to tackle these challenges. The individual adaptive evolution mechanism introduces a dynamic approach to optimize convergence-related and diversity-related variables by allocating computational resources to individuals based on their evolution states. This method effectively balances convergence and diversity in the high-dimensional search space. The regional collaboration mechanism, on the other hand, employs an auxiliary population to explore multiple sub-regions to maintain diversity, guiding the main population towards the constrained Pareto front. By combining these two mechanisms within a two-stage algorithm framework, a new algorithm IAERCEA is proposed. IAERCEA and nine other state-of-the-art algorithms are evaluated on several benchmark suites and three dynamic economic emissions dispatch problems. The results show that IAERCEA has better or competitive performance.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"95 ","pages":"Article 101925"},"PeriodicalIF":8.2,"publicationDate":"2025-04-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143823215","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 dueling double deep Q network assisted cooperative dual-population coevolutionary algorithm for multi-objective combined economic and emission dispatch problems","authors":"Xiaobing Yu ,&nbsp;Yuan Wang ,&nbsp;Wen Zhang","doi":"10.1016/j.swevo.2025.101949","DOIUrl":"10.1016/j.swevo.2025.101949","url":null,"abstract":"<div><div>With the increasing demand for electricity and the awareness of environmental protection, requirements for economic efficiency and controlling environmental impact of the power system are increasing. However, traditional power system scheduling usually focuses on ensuring the stability of the power supply, which neglects the optimization of cost and emissions. Therefore, combined economic and emission dispatch (CEED) problem is proposed to overcome this challenge. Due to nonlinear and nonconvex objective functions and narrow feasible regions, the optimization of multi-objective CEED problem encounters many difficulties. A dueling double deep Q network-assisted cooperative dual-population coevolutionary algorithm (D3QN<img>CDCA) is developed to solve multi-objective CEED problems. The proposed algorithm utilizes D3QN to select operators dynamically and adaptively for two populations of coevolutionary algorithm, thus enhancing its adaptability to different practical constrained multi-objective problems and satisfying search needs of different iteration stages. The introduction of D3QN is to overcome the inherent overestimation of DQN and improve the learning efficiency of the network. To comprehensively evaluate its performance, we tested D3QN<img>CDCA on benchmark function sets and applied it to CEED problems in comparison with other competitive algorithms. Results demonstrate that D3QN<img>CDCA outperforms existing methods with an average IGD+ ranking of 1.4286 and an average HV ranking of 1.321 in benchmark function sets. Meanwhile, the proposed algorithm achieves an average improvement of 23.86 % in 6-unit, 23.12 % in 11-unit and 13.51 % in 14-unit CEED problems. The improvement in solution quality demonstrates the effectiveness of D3QN<img>CDCA in solving high-dimensional multi-objective optimization problems, particularly in CEED, highlighting its potential for real-world energy management applications.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"95 ","pages":"Article 101949"},"PeriodicalIF":8.2,"publicationDate":"2025-04-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143843155","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 configuration of high-dimensional hyperparameters in deep convolutional neural networks for classification assisted by surrogate models","authors":"Rui Zhang ,&nbsp;Yuanyuan Zhang ,&nbsp;Chaoli Sun ,&nbsp;Yanjun Zhang ,&nbsp;Zehua Dong ,&nbsp;Xiaobing Wang","doi":"10.1016/j.swevo.2025.101940","DOIUrl":"10.1016/j.swevo.2025.101940","url":null,"abstract":"<div><div>The rationality of hyperparameter configuration in deep convolutional neural networks for classification directly determines its performance. It is challenging to reduce high computational costs effectively and guarantee performance in the configuration of high-dimensional hyperparameters in deep convolutional neural networks for classification. This paper proposes an efficient configuration method that concerns high-dimensional hyperparameters in deep convolutional neural networks for classification assisted by surrogate models. By designing a progressive accumulation dropout neural network surrogate model (PA-Dropout), the contribution of hyperparameters configurations to multi-performance objectives is dynamically measured and then the contribution is iteratively screened. As a result, the fitting efficiency of the PA-Dropout to the relationship between high-dimensional hyperparametric configurations and the multi-objective performance in deep convolutional neural networks for classification with scarce data is improved. A dual-drive interactive dynamic model management strategy (DDIDMMS) is designed, considering the comprehensive evaluation and adaptive weighting calculation of convergence diversity of high-dimensional hyperparametric configuration individuals. Reliable candidate solutions are provided for real evaluation, thereby improving the update efficiency of PA-Dropout. Finally, an efficient configuration of high-dimensional hyperparameters in deep convolutional neural networks for classification is realized. By using DTLZ and WFG benchmark problems with up to 100 decision variables and 20 targets, as well as practical classification tasks, the superiority and generalization of this method are verified when solving the expensive multi-objective optimization problem of CNN high-dimensional hyperparameter configuration for classification tasks.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"95 ","pages":"Article 101940"},"PeriodicalIF":8.2,"publicationDate":"2025-04-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143815542","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":"Adaptive knowledge-based multi-objective evolutionary algorithm for hybrid flow shop scheduling problems with multiple parallel batch processing stages","authors":"Feige Liu ,&nbsp;Xin Li ,&nbsp;Chao Lu ,&nbsp;Wenyin Gong","doi":"10.1016/j.swevo.2025.101929","DOIUrl":"10.1016/j.swevo.2025.101929","url":null,"abstract":"<div><div>Parallel batch processing machines have extensive applications in the semiconductor manufacturing process. However, the problem models in previous studies regard parallel batch processing as a fixed processing stage in the machining process. This study generalizes the problem model, in which users can arbitrarily set certain stages as parallel batch processing stages according to their needs. A Hybrid Flow Shop Scheduling Problem with Parallel Batch Processing Machines (PBHFSP) is solved in this paper. Furthermore, an Adaptive Knowledge-based Multi-Objective Evolutionary Algorithm (AMOEA/D) is designed to simultaneously optimize both makespan and Total Energy Consumption (TEC). Firstly, a hybrid initialization strategy with heuristic rules based on knowledge of PBHFSP is proposed to generate promising solutions. Secondly, the disjunctive graph model has been established based on the knowledge to find the critical-path of PBHFS. Then, a critical-path based neighborhood search is proposed to enhance the exploitation ability of AMOEA/D. Moreover, the search time is adaptively adjusted based on learning experience from Q-learning and Decay Law. Afterward, to enhance the exploration capability of the algorithm, AMOEA/D designs an improved population updating strategy with a weight vector updating strategy. These strategies rematch individuals with weight vectors, thereby maintaining the diversity of the population. Finally, the proposed algorithm is compared with state-of-the-art algorithms. The experimental results show that the AMOEA/D is superior to the comparison algorithms in solving the PBHFSP.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"95 ","pages":"Article 101929"},"PeriodicalIF":8.2,"publicationDate":"2025-04-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143815543","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":"Auxiliary optimization framework based on scaling transformation matrix for large-scale multi-objective problem","authors":"Yuanyuan Ge,&nbsp;Zhanpeng Wang,&nbsp;Hongyan Wang,&nbsp;Fan Cheng,&nbsp;Lei Zhang","doi":"10.1016/j.swevo.2025.101931","DOIUrl":"10.1016/j.swevo.2025.101931","url":null,"abstract":"<div><div>Large-scale multi-objective optimization problems (LSMOPs) usually have a complex continuous search space, and it is difficult for a single optimization strategy to effectively explore the decision space. Meanwhile, the dimensionality reduction strategy is easy to lose the original data information and cannot be recovered in the optimization process. Therefore, this paper proposes an auxiliary optimization framework based on the scaling transformation matrix (AOF-STM) for solving the LSMOPs, which utilizes the optimization information from low-dimensional auxiliary problems to assist in the optimization of high-dimensional original problems. The construction of the scaling transformation matrix (STM) is based on calculating the similarity of the distribution features between the objective space and decision space, and then effective information sharing between different problems is achieved by STM. Specifically, each element STM (<span><math><mrow><mi>i</mi><mo>,</mo><mi>j</mi></mrow></math></span>) reflects the similarity between the <span><math><mi>i</mi></math></span>-th decision variable (in low-dimensional auxiliary problem) and the <span><math><mi>j</mi></math></span>-th decision variable (in high-dimensional original problem). Based on the proposed scaling transformation matrix STM, the information and experience of the low-dimensional auxiliary problem can be effectively used to guide the learning process of the original problem. Experimental results show that on LSMOPs with the 1000 to 50000 decision variables, AOF-STM shows better performance in terms of convergence and diversity than several state-of-the-art algorithms.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"95 ","pages":"Article 101931"},"PeriodicalIF":8.2,"publicationDate":"2025-04-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143815544","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":"Learning-infused optimization for evolutionary computation","authors":"Kun Bian ,&nbsp;Juntao Zhang ,&nbsp;Hong Han ,&nbsp;Jun Zhou ,&nbsp;Yifei Sun ,&nbsp;Shi Cheng","doi":"10.1016/j.swevo.2025.101930","DOIUrl":"10.1016/j.swevo.2025.101930","url":null,"abstract":"<div><div>Evolutionary computation is a class of meta-heuristic algorithm that mimics the process of biological evolution, utilizing information exchange among individuals in the population to iteratively search for optimal solutions. During the evolutionary process, a substantial amount of data is generated, from which valuable evolutionary information can be extracted to assist the algorithm to evolve in a more effective direction. Additionally, neural networks excel at extracting knowledge from data. Motivated by this, we propose a learning-infused optimization (LIO) framework that employs neural networks to learn the evolutionary processes of the algorithms and extract synthesis patterns from the valuable evolutionary information. These synthesis patterns possess excellent generalizability and effectiveness, guiding the algorithm towards better solutions on the original problems and enabling transfer evolution ability, which can improve the performance of the algorithm on new problems. The LIO framework is applied to various algorithms. Experimental results demonstrate that the synthesis patterns extracted from the CEC14 problems not only guide the evolution of the algorithms towards better solutions on the original problems, but also significantly improve the performance of the algorithms on the CEC17 problems.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"95 ","pages":"Article 101930"},"PeriodicalIF":8.2,"publicationDate":"2025-04-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143799059","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 novel disjunctive-graph-based meta-heuristic approach for multi-objective resource-constrained project scheduling problem with multi-skilled staff","authors":"Min Hu ,&nbsp;Min Zhou ,&nbsp;Zikai Zhang ,&nbsp;Liping Zhang ,&nbsp;Yingli Li","doi":"10.1016/j.swevo.2025.101939","DOIUrl":"10.1016/j.swevo.2025.101939","url":null,"abstract":"<div><div>In the implementation of projects, human resources play a crucial role. The effective assignment of multi-skilled staff among project scheduling can enhance the enterprise competitiveness. Hence, this work addresses the resource-constrained project scheduling problem with multi-skilled staff (MS-RCPSP) to minimize project completion time and total salary cost. A position-based mixed-integer linear programming model, a disjunctive graph model and a novel disjunctive-graph-based objective-guided nearest neighborhood search (DO-NNS) algorithm are proposed. The algorithm includes a resource-oriented encoding, a critical path method-based decoding and a nearest neighborhood search mechanism. By analyzing the disjunctive graph model, this work mines six relational attributes and three properties. Further, this algorithm uses these properties to design three objective-guided neighborhood search operators to enhance its performance. Moreover, the enhanced population update strategy is developed to enhance the quality of Pareto solutions. Finally, the experimental results demonstrate that the improvements are effective and the DO-NNS is superior to five latest multi-objective algorithms in terms of achieving higher-quality Pareto solutions.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"95 ","pages":"Article 101939"},"PeriodicalIF":8.2,"publicationDate":"2025-04-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143860533","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":"Performance evaluation for accelerated and efficient prediction of different regression models aggravated with BPSO for enhancing area efficiency through state encoding in sequential circuits","authors":"Kaushik Khatua,&nbsp;Santanu Chattopadhyay,&nbsp;Anindya Sundar Dhar","doi":"10.1016/j.swevo.2025.101919","DOIUrl":"10.1016/j.swevo.2025.101919","url":null,"abstract":"&lt;div&gt;&lt;div&gt;Minimizing circuit size and cost is a critical challenge in digital design, particularly in Finite State Machine (FSM) synthesis, which is essential for sequential systems. FSMs, implemented as Mealy or Moore machines, play a vital role in embedded systems and communication protocols. However, optimizing FSMs is inherently complex due to the NP-hard State Assignment Problem (SAP), which impacts circuit area, performance, and power efficiency. Traditional methods like KISS and NOVA often struggle with scalability and efficiency, highlighting the need for advanced solutions. To address this, we propose a Binary Particle Swarm Optimization (BPSO) approach integrated with regression-based predictive models, including Linear Regression (LR), K-Nearest Neighbor Regression (KNN), and Support Vector Regression (SVR). By leveraging a dataset of particle populations and their fitness evaluations, the predictive framework replaces computationally intensive cost simulators like ESPRESSO/SIS, significantly reducing runtime while maintaining high accuracy. Experimental results demonstrate that the BPSO-based approach achieves significant area cost reductions, with a 4.9% improvement in two-level optimization and 5.62% in multi-level optimization. The predictive model significantly improves computational efficiency, reducing total run-time by &lt;span&gt;&lt;math&gt;&lt;mrow&gt;&lt;mn&gt;1&lt;/mn&gt;&lt;mo&gt;.&lt;/mo&gt;&lt;mn&gt;2&lt;/mn&gt;&lt;mo&gt;×&lt;/mo&gt;&lt;/mrow&gt;&lt;/math&gt;&lt;/span&gt; to &lt;span&gt;&lt;math&gt;&lt;mrow&gt;&lt;mn&gt;1&lt;/mn&gt;&lt;mo&gt;.&lt;/mo&gt;&lt;mn&gt;89&lt;/mn&gt;&lt;mo&gt;×&lt;/mo&gt;&lt;/mrow&gt;&lt;/math&gt;&lt;/span&gt;, with the highest speedup observed in &lt;em&gt;planet&lt;/em&gt; (&lt;span&gt;&lt;math&gt;&lt;mrow&gt;&lt;mn&gt;1&lt;/mn&gt;&lt;mo&gt;.&lt;/mo&gt;&lt;mn&gt;89&lt;/mn&gt;&lt;mo&gt;×&lt;/mo&gt;&lt;/mrow&gt;&lt;/math&gt;&lt;/span&gt;) and the lowest in &lt;em&gt;dk14&lt;/em&gt; (&lt;span&gt;&lt;math&gt;&lt;mrow&gt;&lt;mn&gt;0&lt;/mn&gt;&lt;mo&gt;.&lt;/mo&gt;&lt;mn&gt;85&lt;/mn&gt;&lt;mo&gt;×&lt;/mo&gt;&lt;/mrow&gt;&lt;/math&gt;&lt;/span&gt;). The model’s accuracy is validated by evaluating key performance metrics for different regression techniques. Support Vector Regression (SVR) achieves the highest prediction accuracy with an &lt;span&gt;&lt;math&gt;&lt;msup&gt;&lt;mrow&gt;&lt;mi&gt;R&lt;/mi&gt;&lt;/mrow&gt;&lt;mrow&gt;&lt;mn&gt;2&lt;/mn&gt;&lt;/mrow&gt;&lt;/msup&gt;&lt;/math&gt;&lt;/span&gt; score of 0.987, outperforming KNN (0.986) and LR (0.973). SVR also exhibits the lowest Mean Absolute Percentage Error (MAPE) of 0.0627, followed by LR (0.081) and KNN (0.091). In terms of Mean Squared Error (MSE), KNN performs best with &lt;span&gt;&lt;math&gt;&lt;mrow&gt;&lt;mn&gt;3&lt;/mn&gt;&lt;mo&gt;.&lt;/mo&gt;&lt;mn&gt;85&lt;/mn&gt;&lt;mo&gt;×&lt;/mo&gt;&lt;mn&gt;1&lt;/mn&gt;&lt;msup&gt;&lt;mrow&gt;&lt;mn&gt;0&lt;/mn&gt;&lt;/mrow&gt;&lt;mrow&gt;&lt;mo&gt;−&lt;/mo&gt;&lt;mn&gt;6&lt;/mn&gt;&lt;/mrow&gt;&lt;/msup&gt;&lt;/mrow&gt;&lt;/math&gt;&lt;/span&gt;, followed by SVR (&lt;span&gt;&lt;math&gt;&lt;mrow&gt;&lt;mn&gt;3&lt;/mn&gt;&lt;mo&gt;.&lt;/mo&gt;&lt;mn&gt;92&lt;/mn&gt;&lt;mo&gt;×&lt;/mo&gt;&lt;mn&gt;1&lt;/mn&gt;&lt;msup&gt;&lt;mrow&gt;&lt;mn&gt;0&lt;/mn&gt;&lt;/mrow&gt;&lt;mrow&gt;&lt;mo&gt;−&lt;/mo&gt;&lt;mn&gt;6&lt;/mn&gt;&lt;/mrow&gt;&lt;/msup&gt;&lt;/mrow&gt;&lt;/math&gt;&lt;/span&gt;) and LR (&lt;span&gt;&lt;math&gt;&lt;mrow&gt;&lt;mn&gt;3&lt;/mn&gt;&lt;mo&gt;.&lt;/mo&gt;&lt;mn&gt;96&lt;/mn&gt;&lt;mo&gt;×&lt;/mo&gt;&lt;mn&gt;1&lt;/mn&gt;&lt;msup&gt;&lt;mrow&gt;&lt;mn&gt;0&lt;/mn&gt;&lt;/mrow&gt;&lt;mrow&gt;&lt;mo&gt;−&lt;/mo&gt;&lt;mn&gt;6&lt;/mn&gt;&lt;/mrow&gt;&lt;/msup&gt;&lt;/mrow&gt;&lt;/math&gt;&lt;/span&gt;). Additionally, SVR demonstrates the lowest Mean Bias Deviation (MBD) of &lt;span&gt;&lt;math&gt;&lt;mrow&gt;&lt;mn&gt;3&lt;/mn&gt;&lt;mo&gt;.&lt;/mo&gt;&lt;mn&gt;62&lt;/mn&gt;&lt;mo&gt;×&lt;/mo&gt;&lt;mn&gt;1&lt;/mn&gt;&lt;msup&gt;&lt;mrow&gt;&lt;mn&gt;0&lt;/mn&gt;&lt;/mrow&gt;&lt;mrow&gt;&lt;mo&gt;−&lt;/mo&gt;&lt;mn&gt;6&lt;/mn&gt;&lt;/mrow","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"95 ","pages":"Article 101919"},"PeriodicalIF":8.2,"publicationDate":"2025-04-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143776731","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 multi-objective evolutionary algorithm for feature selection incorporating dominance-based initialization and duplication analysis","authors":"Chuili Chen ,&nbsp;Xiangjuan Yao ,&nbsp;Dunwei Gong ,&nbsp;Huijie Tu","doi":"10.1016/j.swevo.2025.101914","DOIUrl":"10.1016/j.swevo.2025.101914","url":null,"abstract":"<div><div>The primary objective of feature selection is to reduce the number of features while improving classification performance. Therefore, this problem is typically modeled as a multi-objective optimization problem and can be solved using multi-objective evolutionary algorithms (MOEAs). However, feature selection based on weights derived from preferences may lead to the exclusion of specific features, thereby impacting classification performance. Furthermore, if duplicate individuals are not adequately addressed during the evolutionary process, it may adversely affect the convergence and diversity of the population. In this paper, we propose a multi-objective evolutionary algorithm for feature selection incorporating dominance-based initialization and duplication analysis. To filter features impartially, we transform the correlation issues among features, as well as those between features and labels, into a multi-objective optimization problem by assigning corresponding weights based on their dominance relationships. In addressing the duplication problem within the evolutionary process, the disparity between duplicate individuals as well as between duplicate individuals and elite solutions is analyzed to systematically eliminate redundancy. In the experiments, the proposed method was compared with two classical algorithms and three feature selection algorithms across thirteen datasets. The experimental results indicate that the proposed method exhibits superior classification and optimization performance across the majority of datasets.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"95 ","pages":"Article 101914"},"PeriodicalIF":8.2,"publicationDate":"2025-04-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143746430","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":"Heterogeneous approximation-assisted search for expensive multi-objective optimization","authors":"Shufen Qin,&nbsp;Chaoli Sun","doi":"10.1016/j.swevo.2025.101926","DOIUrl":"10.1016/j.swevo.2025.101926","url":null,"abstract":"<div><div>The cheap surrogate model is commonly used to guide the multi-objective optimization algorithm in the search for the optimum of the expensive optimization problem. However, modeling diversity and its quality are the keys that affect the performance of approximating the original problem. Using multiple heterogeneous models can provide more diverse approximations for complicated optimization problems. Meanwhile, the location relationship between individuals and training samples is a potential benefit for selecting infill individuals to update the model. Therefore, this paper proposes to train two heterogeneous models for each expensive objection function, with the update of the models using the promising individuals based on the approximated domination relationship and the crowding distance between individuals and evaluated samples. Differently, the function estimation of each individual is the sum of two predicted values in a probability-weighted way together with its uncertainty. In addition, the promising individuals are selected by the dominant numbers or the distance to the decision domain center and the crowding distance to the neighbors, otherwise adopting the difference in convergence and crowding distance between all candidates and the training samples to select the individual for expensive function evaluations if the training set dominates all offspring individuals. Experimental studies analyze the effectiveness of the heterogeneous approximation-based guiding search and examine the superiority of the proposed algorithm compared to five recent epidemic optimization algorithms for DTLZ, WFG benchmark problems, and a practical application.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"95 ","pages":"Article 101926"},"PeriodicalIF":8.2,"publicationDate":"2025-04-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143746429","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}