Open AccessA new full waveform analysis approach using simulated tempering Markov chain Monte Carlo method
Author(s) -
Yin Wang,
Weiji He,
Guohua Gu,
Qian Chen
Publication year - 2014
Publication title -
wuli xuebao
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.199
H-Index - 47
ISSN - 1000-3290
DOI - 10.7498/aps.63.164205
Subject(s) - parallel tempering , markov chain monte carlo , computer science , markov chain , waveform , metropolis–hastings algorithm , monte carlo method , algorithm , computation , tempering , convergence (economics) , mathematical optimization , hybrid monte carlo , artificial intelligence , mathematics , machine learning , statistics , bayesian probability , materials science , telecommunications , radar , economics , composite material , economic growth
To reconstruct the target shape distribution in the distance, full waveform analysis algorithm is utilized by extracting and analyzing the number of the peaks, the time of the peak maximum and other parameters. A novel fast full waveform analysis algorithm (simulated tempering Markov chain Monte Carlo algorithm, STMCMC) is proposed, which is able to process the waveform data automatically. For the different types of the parameters, simulated tempering strategy and the Metropolis strategy are presented. In simulated tempering strategy, due to the demand of speed or accuracy, active intervention tempering is used to control the process of solving the vector parameters. On the other hand, the Metropolis strategy is adopted for non-vector parameters to reduce computation amount. Both the strategies are based on Markov chain algorithm, and meanwhile can hold the convergence of the Markov chain, which makes the STMCMC algorithm robust.