Adaptive importance sampling technique for markov chains using stochastic approximation

Ahamed, T. P. I. ; Borkar, V. S. ; Juneja, S. (2006) Adaptive importance sampling technique for markov chains using stochastic approximation Operations Research, 54 (3). pp. 489-504. ISSN 0030-364X

[img]
Preview
PDF - Publisher Version
226kB

Official URL: http://portal.acm.org/citation.cfm?id=1246572

Related URL: http://dx.doi.org/10.1287/opre.1060.0291

Abstract

For a discrete-time finite-state Markov chain, we develop an adaptive importance sampling scheme to estimate the expected total cost before hitting a set of terminal states. This scheme updates the change of measure at every transition using constant or decreasing step-size stochastic approximation. The updates are shown to concentrate asymptotically in a neighborhood of the desired zero-variance estimator. Through simulation experiments on simple Markovian queues, we observe that the proposed technique performs very well in estimating performance measures related to rare events associated with queue lengths exceeding prescribed thresholds. We include performance comparisons of the proposed algorithm with existing adaptive importance sampling algorithms on some examples. We also discuss the extension of the technique to estimate the infinite horizon expected discounted cost and the expected average cost.

Item Type:Article
Source:Copyright of this article belongs to Institute for Operations Research and the Management Sciences.
ID Code:5324
Deposited On:18 Oct 2010 08:41
Last Modified:16 May 2016 15:50

Repository Staff Only: item control page