Publication detail

Sampling strategy for feasible high dimensional Monte Carlo computations

PODROUŽEK, J. BUCHER, C.

Original Title

Sampling strategy for feasible high dimensional Monte Carlo computations

Type

conference paper

Language

English

Original Abstract

The proposed sampling strategy enables feasible computation of high dimensional Monte Carlo simulation tasks by minimizing the number of necessary executions. The original aspect of this contribution is a special sampling design that focuses on realizations of stochastic nonstationary processes as input for computationally intensive models in a seismic protection context. Since nonlinear oscillators always respond in a very uncertain manner to random vibrations, the input and output mapping is based on small sample training and image processing. Application example demonstrates the benefits and limitations of the nontraditional approach and implies application analogies from across various disciplines, such as hydrology, water resources, etc.

Keywords

stochastic process, Critical excitation, Reliability analysis, Importance sampling, Identification problem

Authors

PODROUŽEK, J.; BUCHER, C.

Released

19. 7. 2013

Location

Brno, Czech Republic

ISBN

978-80-214-4800-1

Book

11th International Probabilistic Workshop

Pages from

317

Pages to

323

Pages count

7

URL

server.stm.fce.vutbr.cz

BibTex

@inproceedings{BUT131106,
  author="Jan {Podroužek} and Christian {Bucher}",
  title="Sampling strategy for feasible high dimensional
Monte Carlo computations",
  booktitle="11th International Probabilistic Workshop",
  year="2013",
  pages="317--323",
  address="Brno, Czech Republic",
  isbn="978-80-214-4800-1",
  url="server.stm.fce.vutbr.cz"
}