Swarm and Evolutionary Computation最新文献

{"title":"A high-dimensional feature selection algorithm via fast dimensionality reduction and multi-objective differential evolution","authors":"Xuezhi Yue ,&nbsp;Yihang Liao ,&nbsp;Hu Peng ,&nbsp;Lanlan Kang ,&nbsp;Yuan Zeng","doi":"10.1016/j.swevo.2025.101899","DOIUrl":"10.1016/j.swevo.2025.101899","url":null,"abstract":"<div><div>The multi-objective feature selection problem typically involves two key objectives: minimizing the number of selected features and maximizing classification performance. However, most multi-objective evolutionary algorithms (MOEAs) face challenges in high-dimensional datasets, including low search efficiency and potential loss of search space. To address these challenges, this paper proposes a hybrid algorithm based on fast dimensionality reduction and multi-objective differential evolution with redundant and preference processing (termed DR-RPMODE). In DR-RPMODE, the DR phase uses the freezing and activation operators to remove many irrelevant and redundant features in the high-dimensional datasets, thereby achieving fast dimensionality reduction. Subsequently, the RPMODE algorithm continues the search on the reduced datasets, improving the traditional differential evolutionary framework from two aspects: duplicated and redundant solutions are filtered by redundant handling, and a preference handling method that pays more attention to classification performance is designed for different preference objectives of decision-makers. In the experiment, DR-RPMODE is compared with seven feature selection algorithms on 16 classification datasets. The results indicate that DR-RPMODE outperforms the comparison algorithms on most datasets, demonstrating that it not only achieves outstanding optimization performance but also obtains good classification and scalability results.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"94 ","pages":"Article 101899"},"PeriodicalIF":8.2,"publicationDate":"2025-03-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143551979","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":"EABC-AS: Elite-driven artificial bee colony algorithm with adaptive population scaling","authors":"Ruiyang Lin ,&nbsp;Zesong Xu ,&nbsp;Liyang Yu ,&nbsp;Tongquan Wei","doi":"10.1016/j.swevo.2025.101893","DOIUrl":"10.1016/j.swevo.2025.101893","url":null,"abstract":"<div><div>The Artificial Bee Colony Algorithm (ABC) is a widely recognized optimization algorithm known for its effectiveness. However, many variants of the ABC algorithm fail to fully leverage the potential of each population, and their inherent random search strategies often limit the algorithm’s convergence capabilities, leading to diminished performance. To address these issues, we introduce an enhanced version of the ABC algorithm, which incorporates two essential features: adaptive population scaling and an elite-driven evolutionary strategy. The adaptive population scaling mechanism dynamically adjusts the population size of each bee colony based on their respective function, and the elite-driven evolutionary strategy with external archive makes bees evolve by utilizing information from elite individuals while ensuring diversity is maintained. These two features enhance the algorithm’s convergence ability. We employ the CEC 2017 and CEC 2022 benchmarks to assess the optimization capabilities of the proposed algorithm. The experimental results indicate that the EABC-AS algorithm displays significant competitiveness relative to CEC excellent algorithms and other state-of-the-art (SOTA) algorithms.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"94 ","pages":"Article 101893"},"PeriodicalIF":8.2,"publicationDate":"2025-03-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143534999","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":"Benchmarking footprints of continuous black-box optimization algorithms: Explainable insights into algorithm success and failure","authors":"Ana Nikolikj ,&nbsp;Mario Andrés Muñoz ,&nbsp;Tome Eftimov","doi":"10.1016/j.swevo.2025.101895","DOIUrl":"10.1016/j.swevo.2025.101895","url":null,"abstract":"<div><div>The practices for comparing black-box optimization algorithms based on performance statistics over a benchmark suite are being increasingly criticized. Critics argue that these practices fail to explain why particular algorithms outperform others. Consequently, there is a growing demand for more robust comparison methods that assess the overall efficiency of the algorithms in terms of performance and also consider the specific landscape properties of the optimization problems on which the algorithms are compared. This study introduces a novel approach for comparing algorithms based on the concept of an <em>algorithm footprint</em>, which aims to identify easy and challenging problem instances for a given algorithm. A unique footprint is assigned to each algorithm and then compared, to highlight problem instances where an algorithm either uniquely succeeds or falls, as well as how the algorithms complement each other across the problem instances. Our solution employs a multi-task regression model (MTR) to simultaneously link the performance of multiple algorithms with the landscape features of the problem instances. By applying an Explainable Machine Learning (XML) technique, we quantify and compare the importance of the landscape features for each algorithm. The methodology is applied to a portfolio of three different BBO algorithms, highlighting their success and failure on the Black-Box Optimization Benchmarking (BBOB) suite. The efficacy of our approach is further demonstrated through a comparative analysis with two existing algorithm comparison methods, showcasing the robustness and depth of insights provided by the proposed approach.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"94 ","pages":"Article 101895"},"PeriodicalIF":8.2,"publicationDate":"2025-03-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143551938","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}

{"title":"State-space adaptive exploration for explainable particle swarm optimization","authors":"Mehdi Alimohammadi,&nbsp;Mohammad-R. Akbarzadeh-T","doi":"10.1016/j.swevo.2025.101868","DOIUrl":"10.1016/j.swevo.2025.101868","url":null,"abstract":"<div><div>A systems theory framework for swarm optimization algorithms promises the rigorous analysis of swarm behaviors and systematic approaches that could avoid ad hoc parameter settings and achieve guaranteed performances. However, optimization processes must treat various systems theory concepts, such as stability and controllability, differently, as swarm optimization relies on preserving diversity rather than reaching uniform agent behavior. This work addresses this duality of perspective and proposes State-Space Particle Swarm Optimization (SS-PSO) using the feedback concept in control systems theory. By exploiting the hidden analogy between these two paradigms, we introduce the concept of controllability for optimization purposes through state-space representation. Extending controllability to particle swarm optimization (PSO) highlights the ability to span the search space, emphasizing the significance of particles' movement rather than their loss of diversity. Furthermore, adaptive exploration (AE) using an iterative bisection algorithm is proposed for the PSO parameters that leverages this controllability measure and its minimum singular value to facilitate explainable swarm behaviors and escape local minima. Theoretical and numerical analyses reveal that SS-PSO is only uncontrollable when the cognitive factor is zero, implying less exploration. Finally, AE enhances exploration by increasing the controllability matrix's minimum singular value. This result underscores the profound connection between the controllability matrix and exploration, a critical insight that significantly enhances our understanding of swarm optimization. AE-based State-Space-PSO (AESS-PSO) shows improved exploration and performance over PSO in 86 SOP and CEC benchmarks, particularly for smaller populations.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"94 ","pages":"Article 101868"},"PeriodicalIF":8.2,"publicationDate":"2025-03-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143528710","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":"Multi-objective genetic programming based on decomposition for feature learning in image classification","authors":"Tuo Zhang ,&nbsp;Ying Bi ,&nbsp;Jing Liang ,&nbsp;Bing Xue ,&nbsp;Mengjie Zhang","doi":"10.1016/j.swevo.2025.101875","DOIUrl":"10.1016/j.swevo.2025.101875","url":null,"abstract":"<div><div>Image classification presents a challenge due to its high dimensionality and extensive variations. Feature learning is a powerful method in addressing this challenge, constituting a multi-objective problem aimed at maximizing classification accuracy and minimizing the number of learned features. A few multi-objective genetic programming (MOGP) methods have been proposed to optimize these two objectives, simultaneously. However, existing MOGP methods ignore the characteristics of feature learning tasks. Therefore, this work proposes a decomposition-based MOGP approach with a global replacement strategy for feature learning in data-efficient image classification. To handle the different value ranges of the two objectives, a transformation function is designed to uniform the range of the number of learned features. In addition, a preference-based decomposition strategy is proposed to address the preference for the objective of classification accuracy. The proposed approach is compared with existing MOGP methods for feature learning on five different image classification datasets with different numbers of training images. The experimental results demonstrate the effectiveness of the proposed approach by achieving better HVs than or comparable to the existing MOGP methods in at least 13 out of 20 cases and classification accuracy significantly better than a popular neural architecture search method in all cases.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"94 ","pages":"Article 101875"},"PeriodicalIF":8.2,"publicationDate":"2025-03-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143527381","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":"Landscape features in single-objective continuous optimization: Have we hit a wall in algorithm selection generalization?","authors":"Gjorgjina Cenikj ,&nbsp;Gašper Petelin ,&nbsp;Moritz Seiler ,&nbsp;Nikola Cenikj ,&nbsp;Tome Eftimov","doi":"10.1016/j.swevo.2025.101894","DOIUrl":"10.1016/j.swevo.2025.101894","url":null,"abstract":"<div><div>The process of identifying the most suitable optimization algorithm for a specific problem, referred to as algorithm selection (AS), entails training models that leverage problem landscape features to forecast algorithm performance. A significant challenge in this domain is ensuring that AS models can generalize effectively to novel, unseen problems. This study evaluates the generalizability of AS models based on different problem representations in the context of single-objective continuous optimization. In particular, it considers the most widely used Exploratory Landscape Analysis features, as well as recently proposed Topological Landscape Analysis features, and features based on deep learning, such as DeepELA, TransOptAS and Doe2Vec. Our results indicate that when presented with out-of-distribution evaluation data, none of the feature-based AS models outperform a simple baseline model, i.e., a Single Best Solver.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"94 ","pages":"Article 101894"},"PeriodicalIF":8.2,"publicationDate":"2025-02-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143520855","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}

{"title":"Smart cities optimization using computational intelligence in power-constrained IoT sensor networks","authors":"Khalid A. Darabkh,&nbsp;Muna Al-Akhras","doi":"10.1016/j.swevo.2025.101889","DOIUrl":"10.1016/j.swevo.2025.101889","url":null,"abstract":"<div><div>This paper introduces the Innovative Clustering Energy Efficient Equilibrium Optimizer-based Multi-Hop Routing Protocol (ICEE-EO-MHRP) for addressing the energy constraint in Internet of Things (IoT) network clustering utilizing the Equilibrium Optimizer (EO), a yet efficient computational intelligence method that is used for selecting Designated Cluster Head (DCH) and Backup DCH (BDCH). Additionally, ICEE-EO-MHRP deals with the IoT energy problem by incorporating a novel cost function that ends up of selecting Designated Relays (DRs) and backup DRs for the purpose of forwarding the traffic towards the sink node. Our protocol substantially reduces messages’ exchanges between IoT Sensor Nodes (SNs) by making the replacement of DCH and BDCH dependent on their energy levels dropping below a threshold. To ensure a balanced communication load and efficient scheduling, an innovative deterministic distributed-time division multiple access system is employed. Not only to this extent, but we address data redundancy issue, raised among those quite adjacent SNs, and accordingly propose an efficient management that guarantees having a coherent protocol. In addition to that, device and link failures are professionally addressed by suggesting recovery mechanisms that optimize the proposed protocol. Dealing with these impairments puts our approach well ahead of the competition since it addresses the most practical issues and scenarios, particularly those with challenging environmental constraints. The simulation results demonstrate primarily that our protocol significantly improves the network lifetime by 157.83 % and 109.81 % in comparison to Particle Swarm Optimization and Tabu Search (Tabu-PSO) and Energy-Efficient CH Selection by Improved Sparrow Search Algorithm utilizing Differential Evolution (EECHS-ISSADE), respectively. Comparing ICEE-EO-MHRP to Tabu-PSO and EECHS-ISSADE reveals improvements in residual energy of 335.87 % and 230.05 %, respectively. Furthermore, in comparison to Tabu-PSO and EECHS-ISSADE, the proposed protocol optimizes the throughput by 252.36 % and 168.64 %, respectively. In terms of average delay, our protocol outperforms Tabu-PSO, EECHS-ISSADE, PEGASIS with Artificial Bee Colony (PEG-ABC), Metaheuristics Cluster-based Routing Technique for Energy-Efficient WSN (MHCRT-EEWSN), as well as Hybrid Bald Eagle Search Optimization Algorithm (HBESAOA) by improvements of 57.53 %, 55.15 %, 86.89 %, 20.52 %, and 94.60 %, respectively.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"94 ","pages":"Article 101889"},"PeriodicalIF":8.2,"publicationDate":"2025-02-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143512257","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":"Energy-efficient task scheduling with binary random faults in cloud computing environments","authors":"Lei Jin ,&nbsp;Jie Yuan ,&nbsp;Dequn Zhou ,&nbsp;Xiuzhi Sang ,&nbsp;Shi Chen ,&nbsp;Xianyu Yu ,&nbsp;Guohui Lin","doi":"10.1016/j.swevo.2025.101877","DOIUrl":"10.1016/j.swevo.2025.101877","url":null,"abstract":"<div><div>Fault management and energy consumption control have become focal topics in the rapid development of cloud computing services. This paper addresses the task scheduling problem with binary random faults in the networking and power supply of cloud computing environments and proposes a task scheduling model with the multiobjectives of minimizing energy consumption and task completion time while maximizing task completion rate. An estimation of distribution algorithm (EDA) with crowding distance (C) and neighborhood search (N) (EDA-CN) is designed for the model, into which a multi-model probability matrix, regional dislocation backup mechanism, neighborhood search operator, and crowding distance operator are integrated. Numerical experiments examine the effectiveness of EDA-CN in comparison with EDA, EDA-C, and the classic non-dominated sorting genetic algorithm III (NSGA3). The results show that EDA-CN consistently outperformed EDA and EDAC, and EDA-CN and NSGA3 performed comparably often yet EDA-CN still outperformed statistically significantly.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"94 ","pages":"Article 101877"},"PeriodicalIF":8.2,"publicationDate":"2025-02-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143487028","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":"Online feature subset selection for mining feature streams in big data via incremental learning and evolutionary computation","authors":"Yelleti Vivek ,&nbsp;Vadlamani Ravi ,&nbsp;P. Radha Krishna","doi":"10.1016/j.swevo.2025.101896","DOIUrl":"10.1016/j.swevo.2025.101896","url":null,"abstract":"<div><div>Online streaming feature subset selection (OSFSS) presents a noteworthy challenge when data samples arrive rapidly and in a time-dependent manner. The complexity of this problem is further exacerbated when features arrive as a stream. Despite several attempts to solve OSFSS over feature streams, existing methods lack scalability, cannot handle interaction effects among features, and fail to efficiently handle high-velocity feature streams. To address these challenges, we propose a novel wrapper-for OSFSS named OSFSS-W (wrapper-for OSFSS), specifically designed to mine feature streams within the Apache Spark environment. Our proposed method incorporates (i) two vigilance tests: for removing (a) irrelevant features and (b) redundant features (ii) incremental learning and (iii) a tolerance-based feedback mechanism that retains and utilizes previous knowledge while adhering to the predefined tolerance thresholds. Additionally, for the purpose of optimization, we introduce a Bare Bones Particle Swarm Optimization (BBPSO-L) algorithm driven by the logistic distribution. Further, the BBPSO-L is parallelized within Apache Spark, following an island-based approach. We evaluated the performance of the proposed algorithm on the datasets taken from the cybersecurity, bioinformatics, and finance domains. The results demonstrate that incorporating two vigilance tests coupled with a tolerance-based feedback mechanism significantly improved the median Area under the receiver operating characteristic curve (AUC) and median cardinality across all datasets.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"94 ","pages":"Article 101896"},"PeriodicalIF":8.2,"publicationDate":"2025-02-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143487580","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 learning-based dual-population optimization algorithm for hybrid seru system scheduling with assembly","authors":"Yuting Wu ,&nbsp;Ling Wang ,&nbsp;Rui Li ,&nbsp;Yuxiang Xu ,&nbsp;Jie Zheng","doi":"10.1016/j.swevo.2025.101901","DOIUrl":"10.1016/j.swevo.2025.101901","url":null,"abstract":"<div><div>As the personalized demand increases, the hybrid seru system (HSS) has emerged as an efficient production paradigm to address the volatile market and intricate production conditions due to its reconfigurability. To satisfy the actual production demands, it is common to consider multiple assembly stages in the HSS. However, the increasing complexity poses challenges for the design of scheduling optimization algorithms. In this paper, a learning-based dual-population optimization algorithm (LDPOA) is designed for the hybrid seru system scheduling problem with assembly. Based on a problem-specific decomposition paradigm, a dual-population cooperative search framework is proposed to enhance the exploration capability by focusing on different subproblem optimizations in different populations. During the evolution, a fusion strategy and filtering mechanism are designed to avoid invalid searches by allocating computing resources to more potential individuals. A learning-guided search mode selection strategy and a population communication strategy are proposed to further improve search efficiency and population diversity. Finally, the adjustment strategies are proposed to improve the solution quality by leveraging problem knowledge. Extensive experiments are conducted to assess the performance of the LDPOA. The comparisons show that the HSS can improve production efficiency by 35.3 % compared to the traditional manufacturing mode.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"94 ","pages":"Article 101901"},"PeriodicalIF":8.2,"publicationDate":"2025-02-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143528709","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}