Bayesian Approach to Optimal Release Policy of Software System

HeeSoo KIM  Shigeru YAMADA  DongHo PARK  

IEICE TRANSACTIONS on Fundamentals of Electronics, Communications and Computer Sciences   Vol.E88-A   No.12   pp.3618-3626
Publication Date: 2005/12/01
Online ISSN: 
DOI: 10.1093/ietfec/e88-a.12.3618
Print ISSN: 0916-8508
Type of Manuscript: PAPER
Category: Reliability, Maintainability and Safety Analysis
NHPP,  software reliability growth model,  optimal release policy,  Bayesian approach,  

Full Text: PDF(1.9MB)
>>Buy this Article

In this paper, we propose a new software reliability growth model which is the mixture of two exponential reliability growth models, one of which has the reliability growth and the other one does not have the reliability growth after the software is released upon completion of testing phase. The mixture of two such models is characterized by a weighted factor p, which is the proportion of reliability growth part within the model. Firstly, this paper discusses an optimal software release problem with regard to the expected total software cost incurred during the warranty period under the proposed software reliability growth model, which generalizes Kimura, Toyota and Yamada's (1999) model with consideration of the weighted factor. The second main purpose of this paper is to apply the Bayesian approach to the optimal software release policy by assuming the prior distributions for the unknown parameters contained in the proposed software reliability growth model. Some numerical examples are presented for the purpose of comparing the optimal software release policies depending on the choice of parameters by the non-Bayesian and Bayesian methods.