Estimation Of Complex System Reliability In Geometric Model Based On Masked Data

System is composed of components. So the reliability of components is boundto affect the system’s reliability directly. When the reliability test was carried out onthe system, we need to gather system life data and system failure reason to analyzethe system reliability. However, sometimes due to equipment or record, the systemfailure reason can not be accurately observed in the process of the actual test, that isto say the component which result in system failure can not be determined, but thesystem life data and a collection of components that may caused system failure. Thiskind of data called masked data.Moreover, because of the limit by test cost, only a relatively small sample takepart in the test. At this time, due to less sample information, the Bayes method isnormally used. At present, it is very rare that using the method of Bayes theoryanalyze reliability of general complex systems described by minimum path.In this paper, based on the masked data, and on the assumption that componentslife under the condition of geometric distribution, we study the reliability of generalcomplex systems which described by the minimum path and composed ofcomponents that independent of each other, using the maximum likelihoodestimation method and the Bayes method. The results are follows:First of all, on account of the masked data, we study general complex systemswhich composed of independent components in geometric model and described bythe minimum path, we get the maximum likelihood estimation of parameter, thenobtain the reliability estimation of components and system under the square loss.Finally, under the gradually increase of the truncated sample II, combined withreliability expression of general complex system which components obeyinggeometric distribution with same and different parameters, using Bayes method, weobtain Bayesian estimation and EB estimation of the system reliability under thesquare loss.