Large deviation theory has provided important clues for the choice of importance sampling measures for Monte Carlo evaluation of exceedance probabilities. However, Glasserman and Wang [Ann. Appl. Probab. 7 (1997) 731–746] have given examples in which importance sampling measures that are consistent with large deviations can perform much worse than direct Monte Carlo. We address this problem by using certain mixtures of exponentially twisted measures for importance sampling. Their asymptotic optimality is established by using a new class of likelihood ratio martingales and renewal theory.
@article{1174323253,
author = {Chan, Hock Peng and Lai, Tze Leung},
title = {Efficient importance sampling for Monte Carlo evaluation of exceedance probabilities},
journal = {Ann. Appl. Probab.},
volume = {17},
number = {1},
year = {2007},
pages = { 440-473},
language = {en},
url = {http://dml.mathdoc.fr/item/1174323253}
}
Chan, Hock Peng; Lai, Tze Leung. Efficient importance sampling for Monte Carlo evaluation of exceedance probabilities. Ann. Appl. Probab., Tome 17 (2007) no. 1, pp. 440-473. http://gdmltest.u-ga.fr/item/1174323253/