Chuili Chen , Xiangjuan Yao , Dunwei Gong , Huijie Tu
{"title":"A multi-objective evolutionary algorithm for feature selection incorporating dominance-based initialization and duplication analysis","authors":"Chuili Chen , Xiangjuan Yao , Dunwei Gong , 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, 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}
{"title":"Metaheuristic-based weight optimization for robust deep reinforcement learning in continuous control","authors":"Gwang-Jong Ko , Jaeseok Huh","doi":"10.1016/j.swevo.2025.101920","DOIUrl":"10.1016/j.swevo.2025.101920","url":null,"abstract":"<div><div>In recent studies, the policy-based deep reinforcement learning (DRL) algorithms have exhibited superior performance in addressing continuous control problems, such as machine arms control and robot gait learning. However, these algorithms frequently face challenges inherent in gradient descent-based weight optimization methods, including susceptibility to local optima, slow learning speeds due to saddle points, approximation errors, and suboptimal hyperparameters. This instability leads to significant performance discrepancies among agent instances trained under identical settings, which complicates the practical application of reinforcement learning. To address this, we propose a metaheuristic-based weight optimization framework designed to mitigate learning instability in DRL for continuous control tasks. The proposed framework introduces a two-phase optimization process, where an additional search phase using swarm intelligence algorithms is conducted at the end of the learning phase utilizing DRL. In numerical experiments, the proposed framework demonstrated superior and more stable performance compared to conventional DRL algorithms in robot locomotion tasks.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"95 ","pages":"Article 101920"},"PeriodicalIF":8.2,"publicationDate":"2025-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143746428","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}
Yang-Zhi Li , Wen Liu , Guo-Sheng Xu , Mao-Duo Li , Kai Chen , Shou-Li He
{"title":"Quantum circuit mapping based on discrete particle swarm optimization and deep reinforcement learning","authors":"Yang-Zhi Li , Wen Liu , Guo-Sheng Xu , Mao-Duo Li , Kai Chen , Shou-Li He","doi":"10.1016/j.swevo.2025.101923","DOIUrl":"10.1016/j.swevo.2025.101923","url":null,"abstract":"<div><div>Current quantum devices are constrained by a limited number of physical qubits and their sparse connectivity. When executing logical quantum circuits on these devices, it is necessary to map them into equivalent circuits adhering to these constraints. The initial allocation of logical qubits to physical ones and the strategic addition of SWAP gates to meet connectivity requirements are critical decisions impacting the performance of mapped circuits. To tackle these challenges, this paper presents a novel quantum circuit mapping method that integrates discrete particle swarm optimization and deep reinforcement learning. Initially, a qubit allocation algorithm using discrete particle swarm optimization with a sorting selection strategy quickly maps logical to physical qubits. Then, an enhanced double deep-Q-network based quantum gate scheduling algorithm with an action space search strategy obtains a SWAP addition scheme that results in shallower depth and fewer additional quantum gates. Comparisons on benchmark datasets (B131 and B114) and the IBM Q20 quantum device show that our method outperforms others regarding algorithm runtime, mapped circuit depth, and the number of added SWAP gates. It also demonstrates scalability on large-scale circuits and IBM Q127. Compared to heuristic methods (subgraph isomorphism and filtered depth-limited search based quantum circuit mapping algorithms), it reduces the average number of added SWAP gates by 29.1% and the average runtime by 41.95%. Compared to a recent reinforcement learning method using Monte Carlo Tree Search for mapping, it decreases the average added depth by 26.05% and the average number of added SWAP gates by 25.58%. Moreover, compared to commercial compilers tket and qiskit, the proposed method results in 14.73% and 16.55% fewer SWAP gates, respectively.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"95 ","pages":"Article 101923"},"PeriodicalIF":8.2,"publicationDate":"2025-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143740073","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}
Peng Chen , Jing Liang , Kangjia Qiao , Xuanxuan Ban , P.N. Suganthan , Hongyu Lin , Jilong Zhang
{"title":"A single-objective Sequential Search Assistance-based Multi-Objective Algorithm Framework","authors":"Peng Chen , Jing Liang , Kangjia Qiao , Xuanxuan Ban , P.N. Suganthan , Hongyu Lin , Jilong Zhang","doi":"10.1016/j.swevo.2025.101916","DOIUrl":"10.1016/j.swevo.2025.101916","url":null,"abstract":"<div><div>In recent years, multi-objective optimization has garnered significant attention from researchers. Evolutionary algorithms are proven to be highly effective in solving complex optimization problems in plenty of cases. However, in the pursuit of improved performance, the focus on generality and efficiency has gradually been sidelined. To address this problem, this paper proposes a generalized framework, called Single-objective Sequential Search Assistance-based Multi-objective Algorithm Framework (SSMAF), to enhance the efficiency of existing multi-objective algorithms while reducing computational costs. The framework comprises two phases. The first phase involves two mechanisms to expedite the convergence of the population: (1) A Sequential Search Mechanism (SSM) is utilized to sequentially search corner solutions to enhance the quality of final population, which includes a corner solution search step and a standard solution detection step to search the Pareto Front (PF) while avoiding obtaining unexpected solutions; (2) A Diversity Search Method (DSM) is designed to conduct reinforced searches within localized regions and assess the population’s crowding degree to prevent it from getting stuck in local optima. After obtaining a population with better distribution, the existing multi-objective algorithms can regard it as the initial population to further search the PF. In the experiments, SSMAF is compared with 13 existing algorithms on 42 widely used benchmark test problems and 4 real-world problems. The experimental results show that SSMAF simultaneously improves the solution quality of existing algorithms while reducing their computational complexity.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"95 ","pages":"Article 101916"},"PeriodicalIF":8.2,"publicationDate":"2025-03-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143714599","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}
Shijie Zhao , Tianran Zhang , Lei Zhang , Jinling Song
{"title":"OS-BiTP: Objective sorting-informed bidomain-information transfer prediction for dynamic multiobjective optimization","authors":"Shijie Zhao , Tianran Zhang , Lei Zhang , Jinling Song","doi":"10.1016/j.swevo.2025.101918","DOIUrl":"10.1016/j.swevo.2025.101918","url":null,"abstract":"<div><div>Prediction response mechanisms based on transfer learning are extensively prevalent in dynamic multiobjective optimization algorithms (DMOAs), which transform historical information into a new environment for tracking the Pareto set (PS) or front (PF). However, many existing methods learn information of overall changes from old to new populations for prediction. Due to the different characteristics of individual variation within the population, this inevitably causes the valid information of more relevant individuals to be partially weakened during the training process, thus reducing transfer prediction-based accuracy. Therefore, this paper proposes an objective sorting-informed bidomain-information transfer prediction (OS-BiTP) for the DMOA based on individual objective variation, with the aim of transferring individuals within the same characteristics. The three core components in OS-BiTP are variation-based objective sorting (VOS), bidomain-information transfer within objective space (BiTOS), and bidomain-information transfer within decision space (BiTDS). Specifically, VOS divides the current PF into high- and low-objective variation classes and designs a modified linear prediction mechanism to forecast new environmental objective vectors. Afterward, VOS trains an easy transfer learning model to match old and new environmental individuals with the same objective variation classes to increase the transfer efficiency of individuals. To accurately track dynamic PFs and PSs, BiTOS and BiTDS perform intraclass correlation alignment for the same class of objective vectors and nondominated solutions and fine-tune the predicted objective vectors and solutions based on their variation differences. The numerical results demonstrate the superior performance and application of OS-BiTP via a systematic comparison with seven state-of-the-art DMOAs.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"95 ","pages":"Article 101918"},"PeriodicalIF":8.2,"publicationDate":"2025-03-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143714594","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}
Yufeng Wang , Yong Zhang , Chunyu Xu , Wen Bai , Ke Zheng , Wenyong Dong
{"title":"Decomposition-based dual-population evolutionary algorithm for constrained multi-objective problem","authors":"Yufeng Wang , Yong Zhang , Chunyu Xu , Wen Bai , Ke Zheng , Wenyong Dong","doi":"10.1016/j.swevo.2025.101912","DOIUrl":"10.1016/j.swevo.2025.101912","url":null,"abstract":"<div><div>Constrained multi-objective optimization problems require optimizing and solving multiple objectives while satisfying the constraints. However, in the process of solving this problem, some constraints created infeasible obstacle regions, which led to the neglect of a portion of the constrained Pareto front (CPF). In order to solve this problem, A novel decomposition-based dual-population constrained multi-objective evolutionary algorithm (DD-CMOEA) is proposed. DD-CMOEA adopts a dual population collaborative search strategy, which can quickly find CPF. In the first stage, DD-CMOEA conducts dual population searches on CPF and unconstrained Pareto front (UPF) separately. During the search process, sub-population A uses unconstrained global exploration to obtain information that helps sub-population B jump through infeasible obstacle areas. In the second stage, when the convergence of the sub-population searching for UPF stagnates, the angle-based constraint advantage principle is used for reverse search. It ensures that the searched CPF solution set can be evenly distributed throughout the entire search space. The experimental results on three standard benchmark function suites show that DD-CMOEA outperforms the other six state-of-the-art algorithms in solving constrained multi-objective optimization problems.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"95 ","pages":"Article 101912"},"PeriodicalIF":8.2,"publicationDate":"2025-03-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143697061","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}
Bo Wei , Shanshan Yang , Wentao Zha , Li Deng , Jiangyi Huang , Xiaohui Su , Feng Wang
{"title":"Particle swarm optimization algorithm based on comprehensive scoring framework for high-dimensional feature selection","authors":"Bo Wei , Shanshan Yang , Wentao Zha , Li Deng , Jiangyi Huang , Xiaohui Su , Feng Wang","doi":"10.1016/j.swevo.2025.101915","DOIUrl":"10.1016/j.swevo.2025.101915","url":null,"abstract":"<div><div>Feature selection (FS) plays an important role in data preprocessing. However, with the ever-increasing dimensionality of the dataset, most FS methods based on evolutionary computational (EC) face the challenge of “the dimensionality curse”. To address this challenge, we propose an new particle swarm optimization algorithm based on comprehensive scoring framework (PSO-CSM) for high-dimensional feature selection. First, a piecewise initialization strategy based on feature importance is used to initialize the population, which can help to obtain a diversity population and eliminate some redundant features. Then, a comprehensive scoring mechanism is proposed for screening important features. In this mechanism, a scaling adjustment factor is set to adjust the size of the feature space automatically. As the population continues to evolve, its feature space is further reduced so as to focus on the more promising area. Finally, a general comprehensive scoring framework (CSM) is designed to improve the performance of EC methods in FS task. The proposed PSO-CSM is compared with 10 representative FS algorithms on 18 datasets. The experimental results show that PSO-CSM is highly competitive in solving high-dimensional FS problems.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"95 ","pages":"Article 101915"},"PeriodicalIF":8.2,"publicationDate":"2025-03-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143687667","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":"Swarm-intelligence-based value iteration for optimal regulation of continuous-time nonlinear systems","authors":"Ding Wang, Qinna Hu, Ao Liu, Junfei Qiao","doi":"10.1016/j.swevo.2025.101913","DOIUrl":"10.1016/j.swevo.2025.101913","url":null,"abstract":"<div><div>In this article, a swarm-intelligence-based value iteration (VI) algorithm is constructed to resolve the optimal control issue for continuous-time (CT) nonlinear systems. By leveraging the evolutionary concept of particle swarm optimization (PSO), the challenge of gradient vanishing is effectively overcome compared to traditional adaptive dynamic programming (ADP). Specifically, a PSO-based action network is implemented to perform policy improvement, eliminating the reliance on gradient information. Furthermore, within the ADP framework, the swarm-intelligence-based VI algorithm for CT systems is developed to address the challenges associated with constraints of initial admissible conditions and the difficulty of selecting probing signals in the traditional policy iteration method. The theoretical analysis is provided to show the convergence of the developed VI algorithm and the stability of the closed-loop system, respectively. Finally, under affine and non-affine backgrounds, two simulations are conducted to demonstrate the effectiveness and optimality of the established swarm-intelligence-based VI scheme for CT systems.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"95 ","pages":"Article 101913"},"PeriodicalIF":8.2,"publicationDate":"2025-03-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143687666","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":"Optimizing MapReduce efficiency and reducing complexity with enhanced particle Swarm Optimization (MR-MPSO)","authors":"Chander Diwaker , Vijay Hasanpuri , Yonis Gulzar , Bhanu Sharma","doi":"10.1016/j.swevo.2025.101917","DOIUrl":"10.1016/j.swevo.2025.101917","url":null,"abstract":"<div><div>Big data's explosive growth poses serious data management difficulties, especially given the data's unequal distribution across massive databases. Because of this mismatch, traditional software systems are less effective, which leads to complex and wasteful data processing. We provide MapReduce-Modified Particle Swarm Optimization (MR-MPSO), a unique optimization technique, to tackle this problem. This strategy not only improves the administration of enormous datasets but also tackles the complexity issue of data imbalance. The MR framework is used to handle large-scale data processing tasks, with MR-MPSO driving the map and reducing functions. Our technique combines adaptive inertia weight with Particle Swarm Optimization (PSO) to improve the accuracy and efficiency of optimization for 10 unimodal and multimodal benchmark functions. MR-MPSO outperforms four optimization algorithms—MR K-means, MR bat, MR whale, and regular MR-on measures such as fitness value mean, median, and standard deviation. Furthermore, MR-MPSO regularly enhances throughput and average I/O rate, especially in complex write operations, with gains ranging from 1.4 % to 28.9 % in throughput and 2.1 % to 17.7 % in I/O rate over typical MR approaches.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"95 ","pages":"Article 101917"},"PeriodicalIF":8.2,"publicationDate":"2025-03-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143687668","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}