• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文会议>Reliability and optimization of structural systems >Assessment of MCMC algorithms for subset simulation
【24h】

Assessment of MCMC algorithms for subset simulation

机译：评估用于子集仿真的MCMC算法

获取原文
获取原文并翻译 | 示例
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

The subset simulation is an adaptive simulation method that efficiently solves structural reliability problems with a large number of random variables. The method includes sampling from conditional distributions, which is achieved through Markov Chain Monte Carlo (MCMC) algorithms. This paper investigates the performance of different MCMC algorithms for subset simulation. It is found that most of the MCMC algorithms proposed in the literature, based on the Metropolis-Hastings (M-H) sampler, do not present significant improvements over the component-wise M-H algorithm originally proposed for subset simulation in [Au & Beck, Prob Eng Mech, 16(4): 263-277, 2001]. Based on these findings, a novel approach for MCMC sampling in the standard normal space is introduced, which has the benefit of simplicity. Moreover, it is shown that an optimal scaling of either this new approach or the component-wise M-H algorithm can improve the accuracy of the original algorithm, without the need for additional model evaluations.
机译：子集仿真是一种自适应仿真方法，可以有效解决具有大量随机变量的结构可靠性问题。该方法包括从条件分布中采样，这是通过Markov Chain Monte Carlo（MCMC）算法实现的。本文研究了用于子集仿真的不同MCMC算法的性能。结果发现，文献中提出的大多数基于Metropolis-Hastings（MH）采样器的MCMC算法都没有比最初为[Au＆Beck，Prob Eng Mech，16（4）：263-277，2001]。基于这些发现，介绍了一种在标准法向空间中进行MCMC采样的新颖方法，该方法具有简便性的优点。此外，已表明，此新方法或基于组件的M-H算法的最佳缩放比例可以提高原始算法的准确性，而无需其他模型评估。

著录项

  • 来源
  • 会议地点 Yerevan(AM)
  • 作者

    I. Papaioannou; K. Zwirglmaier; D. Straub;

  • 作者单位

    Engineering Risk Analysis Group, Technische Universitaet Muenchen, Germany;

    Engineering Risk Analysis Group, Technische Universitaet Muenchen, Germany;

    Engineering Risk Analysis Group, Technische Universitaet Muenchen, Germany;

  • 会议组织
  • 原文格式 PDF
  • 正文语种 eng
  • 中图分类
  • 关键词

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号