Int J Performability Eng ›› 2019, Vol. 15 ›› Issue (5): 1352-1359.doi: 10.23940/ijpe.19.05.p11.13521359

Previous Articles     Next Articles

Explore One Factor of Affecting Software Reliability Demonstration Testing Result

Zhenyu Maa,b, Wei Wuc, Wei Zhanga, Jianping Wanga, and Fusheng Liua,*   

  1. a Department of Equipment Support and Remanufacturing, Army Academy of Armored Forces, Beijing, 100072, China
    b Department of Information and Communication, Army Academy of Armored Forces, Beijing, 100072, China
    c Troop No. 63963 of PLA, Beijing, 100071, China
  • Submitted on ;
  • Contact: * E-mail address: 625181316@qq.com

Abstract: When the traditional Bayesian method is used to verify software reliability, prior information cannot be distinguished from experimental information. In this paper, a hybrid Bayesian method is used to verify software reliability. At the same time, the correlation coefficient conception is introduced. This paper determines the correlation coefficient in the hybrid beta distribution respectively through the subjective quantitative method, the goodness of fit method, the Kullback information method, the Spearman correlation coefficient method, and the Kendall correlation coefficient method. The feasibility of five kinds of methods for calculating the correlation coefficients is proven by case analysis, and it is also demonstrated that the distinction between prior information and experimental information has an impact on the acceptable quality level of software reliability demonstration testing.

Key words: software reliability, hybrid Bayesian method, correlation coefficient, demonstration testing