Article
Computer Science, Information Systems
Tianxi Ma, Yunhe Wang, Xiangtao Li
Summary: In this paper, a Convex Combination Multiple Populations Competitive Swarm Optimization algorithm (CDCSO) is proposed to solve the complex search optimization problem of UAVs searching for a moving target. The algorithm combines multiple populations and a novel convex combination update strategy to prevent falling into local optima and improves performance.
INFORMATION SCIENCES
(2023)
Article
Automation & Control Systems
Guang He, Xiao-li Lu
Summary: In this paper, a modified quantum-behaved particle swarm optimization (QDQPSO) algorithm is proposed to solve the problems of low precision and getting trapped in local optima in high-dimensional complex problems. The algorithm utilizes quasi opposite-based learning in the initialization stage to improve search efficiency and convergence speed. The double evolutionary mechanism is applied to update individual locations during the iteration process for enhancing overall performance. Perturbation at global optimum position and bound constraint handling are considered to help escape local optima and maintain population diversity. Experimental results show that QDQPSO performs better in terms of accuracy and stability of optimal solutions compared to other optimization algorithms. Wilcoxon rank-sum test and Friedman test demonstrate significant advantages of the improved algorithm. QDQPSO also exhibits superior performance in solving five practical optimization problems compared to several optimization methods.
ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE
(2023)
Article
Engineering, Multidisciplinary
Jian Zhu, Jianhua Liu, Yuxiang Chen, Xingsi Xue, Shuihua Sun
Summary: The paper introduces the Binary Restructuring Particle Swarm Optimization (BRPSO) algorithm as an adaptation of the Restructuring Particle Swarm Optimization (RPSO) algorithm for solving discrete optimization problems. Unlike other binary metaheuristic algorithms, BRPSO does not use transfer functions, instead relying on comparison results and a novel perturbation term for the particle updating process. The algorithm requires fewer parameters and exhibits high exploration capability, as demonstrated by experiments on feature selection problems.
Article
Construction & Building Technology
Thu Huynh Van, Sawekchai Tangaramvong, Soviphou Muong, Phuc Tran Van
Summary: This paper proposes an enhanced comprehensive learning particle swarm optimization (ECLPSO) method, combined with a Gaussian local search (GLS) technique, for the simultaneous optimal size and shape design of truss structures. The ECLPSO method prevents the premature convergence of local optimal solutions by introducing perturbation-based exploitation and adaptive learning probability techniques, as well as its distinctive diversity of particles. The combination of GLS and ECLPSO results in fast convergence and likelihood to obtain the optimal solution.
Article
Computer Science, Artificial Intelligence
Vipul Mann, Abhishek Sivaram, Laya Das, Venkat Venkatasubramanian
Summary: The study investigates the impact of agent loss on performance in a hostile environment using particle swarm optimization, revealing interesting trade-offs between efficiency, robustness, and overall performance for different network topologies. Small-world networks are observed to perform well under hostile conditions.
SWARM AND EVOLUTIONARY COMPUTATION
(2021)
Article
Computer Science, Artificial Intelligence
G. Castillo-Garcia, L. Moran-Fernandez, V. Bolon-Canedo
Summary: Particle Swarm Optimization is an algorithm that imitates bird flock behavior to find the best solution by exploring search space guided by a fitness function. This study applies the Sticky Binary Particle Swarm Optimization algorithm for feature selection in domain adaptation, comparing its performance with the use of complexity metrics. The results show that although our proposal might slightly degrade classification performance, it is faster and selects fewer features, making it a feasible trade-off.
Article
Computer Science, Artificial Intelligence
Milad Shafipour, Abdolreza Rashno, Sadegh Fadaei
Summary: This paper introduces a feature selection method based on particle distance and feature ranking, which is mathematically proven and experimentally supported to outperform existing methods in multiple evaluation metrics.
EXPERT SYSTEMS WITH APPLICATIONS
(2021)
Article
Computer Science, Artificial Intelligence
Xu Chen, Hugo Tianfield, Wenli Du
Summary: This paper introduces a novel bee-foraging learning PSO (BFL-PSO) algorithm with three different search phases, showing very competitive performance in terms of solution accuracy.
APPLIED SOFT COMPUTING
(2021)
Article
Computer Science, Information Systems
Md. Sakir Hossain, Naim Hasan, Md. Abdus Samad, Hossain M. D. Shakhawat, Joydeep Karmoker, Foysol Ahmed, K. F. M. Nafiz Fuad, Kwonhue Choi
Summary: This paper proposes a new Android ransomware detection method based on traffic analysis, which utilizes particle swarm optimization (PSO) for selecting traffic characteristics and decision tree and random forest classifiers for data traffic classification. The method significantly improves the detection accuracy and achieves high performance in detecting ransomware and its types.
Article
Engineering, Electrical & Electronic
Yu Zhou, Lin Gao, Dong Wang, Wenhui Wu, Zhiqiang Zhou, Tingqun Ye
Summary: In this study, an improved localized feature selection method based on multiobjective binary particle swarm optimization was proposed to address fault diagnosis by utilizing the local distribution of data without the need for balancing strategies.
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT
(2023)
Article
Computer Science, Information Systems
Wafa Shafqat, Sehrish Malik, Kyu-Tae Lee, Do-Hyeun Kim
Summary: This study introduces an optimized-ensemble model for smart home energy consumption management, utilizing ensemble learning and particle swarm optimization (PSO) to tune hyper-parameters and features, resulting in improved prediction accuracy.
Article
Computer Science, Artificial Intelligence
Zi-Jia Wang, Qiang Yang, Yu -Hui Zhang, Shu-Hong Chen, Yuan -Gen Wang
Summary: This article proposes a superiority combination learning distributed particle swarm optimization (SCLDPSO) for large-scale optimization problems. The algorithm adopts a master-slave multi-subpopulation distributed model, achieving full communication and information exchange among different subpopulations, and enhancing diversity. The superiority combination learning strategy significantly improves the worse particle by learning from well-performance subpopulations. Experimental results demonstrate the superiority of SCLDPSO over other state-of-the-art large-scale optimization algorithms.
APPLIED SOFT COMPUTING
(2023)
Article
Computer Science, Artificial Intelligence
An-Da Li, Bing Xue, Mengjie Zhang
Summary: This paper proposes an improved sticky binary PSO algorithm for feature selection problems, which aims to enhance evolutionary performance through new mechanisms such as an initialization strategy, dynamic bits masking, and genetic operations. Experimental results show that ISBPSO achieves higher accuracy with fewer features and reduces computation time compared to benchmark PSO-based FS methods.
APPLIED SOFT COMPUTING
(2021)
Article
Medicine, General & Internal
Abdelghani Dahou, Ahmad O. Aseeri, Alhassan Mabrouk, Rehab Ali Ibrahim, Mohammed Azmi Al-Betar, Mohamed Abd Elaziz
Summary: In this paper, a robust skin cancer detection framework is proposed to improve the accuracy by extracting and learning relevant image representations using a MobileNetV3 architecture. The modified Hunger Games Search (HGS) based on Particle Swarm Optimization (PSO) and Dynamic-Opposite Learning (DOLHGS) is used as a novel feature selection to maximize the model's performance. Experimental results show that the proposed approach outperforms other well-known algorithms in terms of classification accuracy and optimized features.
Article
Physics, Multidisciplinary
Bozena Borowska
Summary: This study proposes a learning competitive swarm optimization algorithm (LCSO) based on particle swarm optimization method and competition mechanism, which improves the search process and achieves higher efficiency compared to other tested methods.