Title :
An introduction to ensemble-average importance sampling of Markov chains
Author_Institution :
Digital Equipment Corp., Marlborough, MA, USA
Abstract :
Using a simple Markov chain as an example, the author introduces a novel single-sample-path-based estimation method, ensemble-average importance sampling (EAIS). The EAIS is shown to be much more efficient than the time-average likelihood-ratio method and has less variance. It does not resort to regenerative structure
Keywords :
Markov processes; estimation theory; Markov chains; ensemble-average importance sampling; time-average likelihood-ratio method; Monte Carlo methods; State estimation; Steady-state;
Conference_Titel :
Decision and Control, 1991., Proceedings of the 30th IEEE Conference on
Conference_Location :
Brighton
Print_ISBN :
0-7803-0450-0
DOI :
10.1109/CDC.1991.261530
