軟體可靠度的新模型與其應用

A New Model for Software Reliability and Its Application

陳瑞河(Ruei-He Chen)

指導教授 ： 許文郁

國立清華大學/理學院/統計學研究所/碩士(2017年)

摘要

軟體可靠度(Software Reliability)分析中，如何建模(Modeling)是軟體工程的重要議題。Goel-Okumoto NHPP Model是常用的模型，Cai, Hu and Bai et al.(2008)指出，該模型不能準確地描述，執行除錯(Debugging)過程中錯誤發生的行為。然而，他們並沒提出修正方法。本論文主要工作之一是，提出一個新的模型，大大改善Goel-Okumoto NHPP Model的缺點，並用電腦模擬比較新舊模型的優劣。

　　發展新的套裝軟體，除錯是重要的步驟之一。如何估計除錯後，軟體的存留錯誤率(Remaining Error Rate)是重要工作。本論文的另一工作是，應用新的模型，我們開發出一個方法，可以用來估計除錯後，軟體的存留錯誤率。最後我們以模擬的方式，比較新的估計方法與現有的估計方法的表現。

關鍵字

軟體可靠度 ； 錯誤率 ； 除錯

並列摘要

In software reliability analysis, modeling is a key step to software engineering. A commonly used class of reliability models is the Goel-Okumoto Non-homogeneous Poisson Process (NHPP) models. Cai, Hu and Bai et al. (2008) pointed out that the Goel-Okumoto NHPP models could not properly describe the behavior of cumulative number of observed software failures. However they did not suggest how to amend the models. In this thesis, we propose a new model that can largely enhance the performance in data fitting. Comparisons are made by computer simulations, and it is found that the new model works much better than the old ones.

In the development of a new software package, debugging is indispensable. Releasing a newly developed software with bugs could be detrimental. How to estimate the remaining error rate of a newly developed software after debugging is an important issue. Our new model can be used to develop a new method for estimating the remaining error rate. The performance of this new method is investigated by computer simulations.

並列關鍵字

Goel-Okumoto NHPP Model ； Error Rate ； Debugging ； Software Reliability

參考文獻

[3]. Shooman, M.L. (1972). Probabilistic models for software reliability prediction, in: Proc. the Fault-Tolerant Computing Symposium, pp. 211–215.

[5]. Littlewood, B. and Verrall, J. (1973). A Bayesian reliability growth model for computer software, Applied Statistics 22 (3), pp. 332–346.

[6]. Musa, J.D. (1975) A theory of software reliability and its application, IEEE Transactions on Software Engineering SE-1 (3), pp. 312–327.

[7]. Halstead, M.H. (1977). Elements of Software Science, North-Holland, Elsevier Science Inc, New York.

[9]. Cai, K.Y. (2000). Towards a conceptual framework of software run reliability modeling, Information Sciences 126, pp. 137–163.