Blending-Based Stochastic Simulator

Jean-Laurent Mallet and A. Shtuka. ( 2000 )
in: Mathematical Geology, 32:3 (367-379)

Abstract

This paper presents a new method of constructing random functions whose realizations can be evaluated efficiently. The basic idea is to "blend", both stochastically and linearly, a limited set of independent initial realizations previously generated by any chosen simulation method. The blending stochastic coefficients are determined in such a way that the new random function so generated has the same mean and covariance functions as the random function used for generating the initial realizations.

Download / Links

BibTeX Reference

@article{mallet:inria-00099365,
 abstract = {This paper presents a new method of constructing random functions whose realizations can be evaluated efficiently. The basic idea is to "blend", both stochastically and linearly, a limited set of independent initial realizations previously generated by any chosen simulation method. The blending stochastic coefficients are determined in such a way that the new random function so generated has the same mean and covariance functions as the random function used for generating the initial realizations.},
 author = {Mallet, Jean-Laurent and Shtuka, A.},
 hal_id = {inria-00099365},
 hal_local_reference = {A00-R-114 || mallet00a},
 hal_version = {v1},
 journal = {{Mathematical Geology}},
 keywords = {geostatistics ; random functions ; simulation},
 note = {Article dans revue scientifique avec comit{\'e} de lecture.},
 number = {3},
 pages = {367-379},
 publisher = {{Springer Verlag}},
 title = {{Blending-Based Stochastic Simulator}},
 url = {https://inria.hal.science/inria-00099365},
 volume = {32},
 year = {2000}
}