A Parametric Empirical Bayesian Software Reliability Model
Volume 5, Number 3, April 2009 - Paper 7 - pp. 259 -266
DAMODARAN DURAISWAMY1 and GOPAL GOVINDASAMY21 Centre for Reliability, STQC Directorate, Ministry of Communication & Information Technology, Government of India, Dr.VSI Estate, Thiruvanmiyur, Chennai-600041 2 Department of Statistics, University of Madras, Chepauk, Chennai-600005.
(Received on April 19, 2008, revised on November 10, 2008)
In this paper, a new parametric empirical Bayesian software reliability model is presented. Times between failures follow generalised exponential distribution with stochastically decreasing order on the failure rate functions of successive failure time intervals with the software tester's intention to improve the software quality by the correction of each failure. With the Bayesian approach, the predictive distribution has been arrived at by combining generalised exponential time between failures and gamma prior distribution for the parameter namely failure rate. The expected time between failure measure, reliability function etc. have been obtained. The posterior distribution of the failure rate measure has been deduced and the mean failure rate is also obtained. For the parameter estimation, least square estimation method has been adopted. The proposed model has been applied to three sets of actual software failure data. It has been observed that the predicted failure times as per the proposed model are closer to the actual failure times. Sum of square errors criteria has been used for comparing the actual time between failure times and predicted time between failures.
Click here to download the paper.
Please note : You will need Adobe Acrobat viewer to view the full articles.