Advances and Applications in Statistics
Volume 39, Issue 2, Pages 149 - 160
(April 2014)
|
|
A SURVEILLANCE SCHEME FOR RELIABILITY GROWTH OF REPAIRABLE SYSTEMS AND SOFTWARE BASED ON NHPP
M. S. Aminzadeh
|
Abstract: In this article, it is assumed that observed failure times of a repairable system or software are realization of a non-homogenous Poisson process (NHPP), A surveillance scheme for monitoring a NHPP failure times under Power Law and Log-logistic intensity models is presented. The proposed method detects whether the elapsed time between two consecutive arrival times is large enough based on a pre-specified intensity level for the process. The implication of large enough elapsed time is a “better reliability”. Numerical examples for illustration purposes are provided based on real data sets. A Mathematica code for computation of MLE, the lower limit for failure times, and the conditional expected failure times based on both Power Law and Log-logistic models is available from the author. |
Keywords and phrases: maximum likelihood estimate, power law intensity, log-logistic intensity, NHPP, reliability growth. |
|
Number of Downloads: 344 | Number of Views: 858 |
|