Title :
Discrete equations and software reliability growth models
Author :
Satoh, Daisuke ; Yamada, Shigeru
Author_Institution :
Service Integration Labs., NTT, Tokyo, Japan
Abstract :
We propose a criterion to determine a more appropriate software reliability growth model in the early development phase. The criterion with a discrete software reliability growth model determines the absolute worth of a model because the discrete software reliability growth model perfectly reproduces the parameter estimates when the data are used as an exact solution of the equation. The values of the proposed criterion of a discrete software reliability growth model are smaller than those of another software reliability growth model in all periods during the test phase with actual data sets. The discrete software reliability growth models are described with difference equations which have exact solutions. The models yield accurate parameter estimates in spite of a small amount of input data in an actual software testing. Therefore, the criterion and discrete software reliability growth models enable us to predict in the early development phase when software can be released.
Keywords :
data analysis; difference equations; discrete systems; parameter estimation; program testing; software reliability; absolute worth; data sets; difference equations; discrete equations; discrete software reliability growth model; early development phase; exact solutions; input data; parameter estimates; software release; test phase; Application software; Costs; Difference equations; Differential equations; Engineering management; Parameter estimation; Phase estimation; Reliability engineering; Software reliability; Software testing;
Conference_Titel :
Software Reliability Engineering, 2001. ISSRE 2001. Proceedings. 12th International Symposium on
Print_ISBN :
0-7695-1306-9
DOI :
10.1109/ISSRE.2001.989471