Scholarlink

<p>The accuracy, efficiency, stable results are major issues of the structural reliability methods. In this paper, the performance-based accuracy and efficiency with robust searching scheme for determining the most probable point (MPP) is proposed by hybrid strategies given by active support vector regression (ASVR) coupled by population-based global best particle swarm optimization (GPSO). The GPSO is proposed based on a random local search of particles for adjusting the global-best position. The GPSO is applied for searching the optimal hyper parameters of SVR model applied for approximating failure domain. The robustness of FORM is enhanced based on the proposed GPSO method while efficiency of reliability method is improved by the applying an advanced machine learning given from the hybrid methods of SVR and MPP search. The accuracy of the reliability analysis for approximating failure domain is enhanced based on hybrid ASVR and GPSO. The hybrid ASVR coupled with PSO and GPSO are compared for accuracy, robustness and efficiency with traditional analytical algorithms and soft computing MPP search of PSO and GPSO. The results indicated that the hybrid ASVR methods -based GPSO are strongly provided more accurate results than traditional methods.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

ASVR-GPSO: a novel hybrid active support vector regression and global-best partial swarm optimization for structural reliability analysis

  • 期刊论文

  • 发表日期：

    2025年9月20日

查看全文
  • Shiyuan Yang,
  • Debiao Meng,
  • Behrooz Keshtegar,
  • Abílio M. P. De Jesus,
  • Mahmoud Alfouneh,
  • Shun-Peng Zhu

摘要

The accuracy, efficiency, stable results are major issues of the structural reliability methods. In this paper, the performance-based accuracy and efficiency with robust searching scheme for determining the most probable point (MPP) is proposed by hybrid strategies given by active support vector regression (ASVR) coupled by population-based global best particle swarm optimization (GPSO). The GPSO is proposed based on a random local search of particles for adjusting the global-best position. The GPSO is applied for searching the optimal hyper parameters of SVR model applied for approximating failure domain. The robustness of FORM is enhanced based on the proposed GPSO method while efficiency of reliability method is improved by the applying an advanced machine learning given from the hybrid methods of SVR and MPP search. The accuracy of the reliability analysis for approximating failure domain is enhanced based on hybrid ASVR and GPSO. The hybrid ASVR coupled with PSO and GPSO are compared for accuracy, robustness and efficiency with traditional analytical algorithms and soft computing MPP search of PSO and GPSO. The results indicated that the hybrid ASVR methods -based GPSO are strongly provided more accurate results than traditional methods.