|
Track 1 Session 10
9:10 to 10:10 a.m. Thursday June 19, 2008
State-of-the-Art Bayesian Reliability
Analysis Using WinBUGS
Bayesian approaches to reliability
analysis have, in the past, been hampered by numerical difficulties
associated with the integral in the denominator of Bayes’ Theorem.
Markov Chain Monte Carlo (MCMC) sampling has provided a
revolutionary solution to this problem, and has made previously
intractable problems easily solvable. The open-source software tool
WinBUGS has made this powerful approach available to the average practitioner. The Bayesian
approach allows utilization of information other than that present
in direct empirical data, makes missing-data and censoring
problems straightforward and allows for model-checking, an
important and often-overlooked step in statistical inference of any
sort. Through examples, this presentation covers the basics of
Bayesian reliability analysis with WinBUGS, from basic parameter
estimation with complete samples to more advanced topics such as
censoring, modeling failure with repair and model validation.
Key Words: Bayesian
Reliability Analysis, WinBUGS, Markov Chain Monte Carlo
|
