| Article ID: | iaor1992934 |
| Country: | Japan |
| Volume: | J74-D-I |
| Issue: | 3 |
| Start Page Number: | 240 |
| End Page Number: | 248 |
| Publication Date: | Mar 1991 |
| Journal: | Transactions of the Institute of Electronics, Information and Communication Engineers |
| Authors: | Osaki Shunji, Yamada Shigeru, Tanio Yoshikazu |
| Keywords: | engineering, statistics: regression, statistics: general |
This paper discusses several methods of software reliability assessment, in which the software failure occurrence phenomenon during testing and operation phases are described by stochastic models. The failure occurrence (or software fault detection) phenomenon during testing phase is formulated by software reliability growth models based on a nonhomogeneous Poisson process. The failure occurrence phenomenon during operation phase is formulated by three methods dependent on the number of remaining software faults at the initial time point of operational use: Environment factor, hazard rate, and cumulative failure rate. Then, we can obtain several operational software reliability measures. Finally, numerical examples of software reliability assessment during operation phase are presented for illustration. [In Japanese.]