Title :
The research of experiment design based on simulation information reliability and Bayesian posteriori risk
Author :
Meng Jing ; Ma Guoqiang
Author_Institution :
Key Lab. of Nat. Defense Sci. & Technol. for Underwater Acoust. Warfare, Zhanjiang, China
Abstract :
In order to carry out safe-or-failure inspection in weaponry test, this article discusses extensively reliability of prior information firstly. Then based on the second category of the maximum likelihood method in choosing prior distribution and Bayesian statistical validation method, the live test program under different simulation data is designed and compared. The example demonstrates that program presented in this article is better and more robust compared with classical experimental design and normal Bayesian method, which can make the best of simulation information.
Keywords :
Bayes methods; inspection; maximum likelihood estimation; military computing; reliability; risk analysis; statistical distributions; underwater equipment; weapons; Bayesian posteriori risk; Bayesian statistical validation method; experiment design; live test program; maximum likelihood method; prior distribution; safe-or-failure inspection; simulation data; simulation information reliability; underwater weaponry equipment test; weaponry test; Bayes methods; Data models; Estimation; Indexes; Inspection; Reliability; Weapons; Bayesian method; Experimental Design; Simulation test; the second category of the maximum likelihood(ML_II);
Conference_Titel :
Natural Computation (ICNC), 2013 Ninth International Conference on
Conference_Location :
Shenyang
DOI :
10.1109/ICNC.2013.6818282
