DocumentCode
597477
Title
On error rates in rare event simulation with heavy tails
Author
Asmussen, S. ; Kortschak, D.
Author_Institution
Dept. of Math., Aarhus Univ., Aarhus, Denmark
fYear
2012
fDate
9-12 Dec. 2012
Firstpage
1
Lastpage
11
Abstract
For estimating P(Sn > x) by simulation where Sk = Y1+...+Yk with Y1, ..., Yn are non-negative and heavy-tailed with distribution F, (Asmussen and Kroese 2006) suggested the estimator nF(Mn-1 V(x - Sn-1)) where Mk = max(Y1, ..., Yk). The estimator has shown to perform excellently in practice and has also nice theoretical properties. In particular, (Hartinger and Kortschak 2009) showed that the relative error goes to 0 as × → ∞. We identify here the exact rate of decay and propose some related estimators with even faster rates.
Keywords
error statistics; estimation theory; exponential distribution; simulation; decay rate; error rate; estimator; heavy tailed distribution; rare event simulation; relative error; subexponential distribution; Educational institutions; Error analysis; Heuristic algorithms; Mathematical model; Monte Carlo methods; Taylor series; Tin;
fLanguage
English
Publisher
ieee
Conference_Titel
Simulation Conference (WSC), Proceedings of the 2012 Winter
Conference_Location
Berlin
ISSN
0891-7736
Print_ISBN
978-1-4673-4779-2
Electronic_ISBN
0891-7736
Type
conf
DOI
10.1109/WSC.2012.6465288
Filename
6465288
Link To Document