使軟體可靠度模型之參數評估更有效且快速的基因演算法

A More Effective and Faster Genetic Algorithm for Parameter Estimation of Software Reliability Growth Models

陳贊元(Tsan-Yuan Chen)

指導教授 ： 黃慶育

國立清華大學/電機資訊學院/資訊工程學系/碩士(2007年)

摘要

為了確保軟體品質和評估軟體可靠度，其中一種方法是使用軟體可靠度成長模型。軟體可靠度成長模型可以用來描述時間與錯誤的關係。錯誤減少因子是淨錯誤減少的比例，是所移除的錯減去新帶入的錯誤除以程式所有的錯。除錯有可能會帶入新的錯誤，而這種現像稱為不完美除錯。此現象會引響錯誤減少因子。而使用此因子的軟體可靠度成長模型並不多。軟體可靠度成長模型中的參數是未知的，必須經由收集到的錯誤資料來進行參數評估。已有許多種參數評估的方法被提出，比如說最小平方法和最大近似估計法。但這些方法都有必需要模型的導函數存在的限制。基因演算法提供了較有彈性解法。在這篇論文中，我們提出了一個改進的基因演算法與考慮錯誤減少因子的軟體可靠度成長模型。而一般軟體可靠度成長模型的參數是使用最大近似估計法和最小平方法來求得，但在這篇論文中，我們使用改進後的基因演算法來求得參數。實驗結果顯示我們提出來的改進的基因演算法比傳統的基因演算法有效率，而我們提出來的軟體可靠度成長模型有著準確的預測能力。

關鍵字

軟體可靠度成長模型 ； 基因演算法

並列摘要

In order to assure software quality and to assess software reliability, one of the current methods is to apply a Software Reliability Growth Model (SRGM). SRGMs can be used to describe software failures as a random process, which is characterized by either times of failures or by the number of failures at fixed times. The fault reduction factor is the ratio of the net fault reduction. That is the removed fault subtracting from the introduced fault per failure. Debugging may introduce a new fault which phenomenon is called imperfect debugging. And this phenomenon may influence the fault reduction factor. Not a lot of SRGMs use this factor. The parameters of SRGMs are unknown and have to be estimated based on collected real software failure data. Several estimation methods have been proposed, like Least Square Estimation (LSE) and Maximum Likelihood Estimation (MLE), but most of them have restrictions such as the existence of derivatives on evaluation functions. On the other hand, Genetic Algorithms (GA) provide us with robust optimization methods in many fields. In this thesis, we propose adding a modified Genetic Algorithm to the parameter estimation of SRGMs and to the SRGMs considering fault reduction factors. Besides, the parameters of SRGMs are usually estimated by LSE and MLE. But we use the modified Genetic Algorithm for parameter estimation. Experimental results show that the proposed Genetic Algorithm is faster and more effective than other traditional Genetic Algorithms and our proposed models can predict the Software reliability more accurately.

參考文獻

[2] M. R. Lyu, Handbook of Software Reliability Engineering, McGraw Hill, 1996.

[3] C. Y. Huang, M. R. Lyu, and S. Y. Kuo, “A Unified Scheme of Some Non-Homogenous Poisson Process Models for Software Reliability Estimation,” IEEE Transactions on Software Engineering, Vol. 29, No. 3, pp. 261-269, 2003.

[4] C. Y. Huang and M. R. Lyu, “Optimal Testing Resource Allocation and Sensitivity Analysis in Software Development,” IEEE Transactions on Reliability, Vol. 54, No. 4, pp. 592-603, 2005.

[5] H. Pham, Software Reliability, Springer-Verlag, 2000.

[7] P. E. Goldberg, Genetic Algorithm in Search Optimization and Machine Learning, New York: Addison-Wesley, 1989.