State-dependent Importance Sampling and large Deviations

Jose Javier Blanchet; Jingchen  Liu; Peter  Glynn

1st International ICST Conference on Performance Evaluation Methodologies and Tools

Research Article

State-dependent Importance Sampling and large Deviations

Cite: BibTeX Plain Text

@INPROCEEDINGS{10.1145/1190095.1190120,
    author={Jose Javier Blanchet and Jingchen  Liu and Peter  Glynn},
    title={State-dependent Importance Sampling and large Deviations},
    proceedings={1st International ICST Conference on Performance Evaluation Methodologies and Tools},
    publisher={ACM},
    proceedings_a={VALUETOOLS},
    year={2012},
    month={4},
    keywords={},
    doi={10.1145/1190095.1190120}
}

Jose Javier Blanchet
Jingchen Liu
Peter Glynn
Year: 2012
State-dependent Importance Sampling and large Deviations
VALUETOOLS
ACM
DOI: 10.1145/1190095.1190120

Jose Javier Blanchet¹^,, Jingchen Liu¹, Peter Glynn²

1: Harvard University
2: Stanford University

Abstract

Large deviations analysis for light-tailed systems provides an asymptotic description of the optimal importance sampler in the scaling of the Law of Large Numbers. As we will show by means of a simple example related to computational finance, such asymptotic description can be interpreted indifferent ways suggesting several importance sampling algorithms, some of them state-dependent. In turn, the performance of the suggested algorithms can be substantially different.

Published: 2012-04-06
Publisher: ACM

: http://dx.doi.org/10.1145/1190095.1190120

State-dependent Importance Sampling and large Deviations

Abstract

About EAI

Community

Publish with EAI