Kernel Method in Importance Sampling Density Estimationby G. L. Ang, Univ of Illinois at Urbana-Champaign, United States,
A. H. -S. Ang, Univ of Illinois at Urbana-Champaign, United States,
W. H. Tang, Univ of Illinois at Urbana-Champaign, United States,
Abstract: The use of the kernel method to estimate the optimal importance sampling density is introduced. This method requires samples generated from an initial direct Monte Carlo run to construct the density estimator. A second set of samples is then generated from the density estimator to estimate the probability of failure. The proposed kernel method is conceptually superior to the use of the normal density in the current application of importance sampling. A significantly lower variance in the estimate of the probability of failure is demonstrated with an example.
Subject Headings: Structural analysis | Probability | Failure analysis | Structural reliability | Density currents | Structural failures | Monte Carlo method | Europe | Monaco | Monte Carlo
Services: Buy this book/Buy this article
Return to search