Title :
Combined Flat Histogram Monte Carlo Method for Efficient Simulation of Communication Systems
Author :
Wijesinghe, P. ; Gunawardana, U. ; Liyanapathirana, R.
Author_Institution :
Sch. of Comput., Eng. & Math., Univ. of Western Sydney, Sydney, NSW, Australia
fDate :
1/1/2012 12:00:00 AM
Abstract :
A modified Monte Carlo (MC) method for efficient sampling of system outputs is derived by combining transition matrix Monte Carlo (TMMC) and Wang-Landau (WL) flat histogram methods. The proposed combination updates the estimation of the WL method using an intermediate estimate derived from the TMMC dynamics such that the variance of the resultant estimate is minimized. Combining the benefits of both methods, the new algorithm can estimate very low probabilities that are not normally reachable by the conventional MC using a realizable sample size.
Keywords :
Monte Carlo methods; matrix algebra; probability; sampling methods; TMMC method; WL method; Wang-Landau flat histogram methods; communication system simulation; intermediate estimation; transition matrix Monte Carlo flat histogram methods; Estimation; Heuristic algorithms; Histograms; MIMO; Markov processes; Monte Carlo methods; Markov chain Monte Carlo; Wang-Landau algorithm; adaptive importance sampling; transition matrix Monte Carlo;
Journal_Title :
Communications Letters, IEEE
DOI :
10.1109/LCOMM.2011.110711.110825
