NHPP-based software reliability assessment using wavelets
NHPP-based software reliability assessment using wavelets
复制标题
基于NHPP的小波软件可靠性评估
DOI：
10.1142/9789813224506_0002
复制
发表时间：
2017
期刊：
影响因子：
--
通讯作者：
X. Xiao and T. Dohi
中科院分区：
文献类型：
--
作者：
Xiao Xiao;X. Xiao and T. Dohi
软件可靠性的定量评估是软件工程中的主要问题之一。在过去的四十年里，研究人员研究了各种定量方法来支持实际软件开发过程中的决策。已经提出了一系列软件可靠性模型(SRM)[Lyu(1996)；Musa等人(1987)；Pham(2000)]，它们采用不同的方法来估计定量软件可靠性，其定义为软件系统在指定的运行期间不会发生故障的概率。其中，基于非齐次泊松过程(NHPP)的软件可靠性模型被广泛应用于软件测试阶段，用于评估软件可靠性、识别软件中剩余故障的数量、确定软件发布时间表。均值函数是一个唯一的参数，它支持基于NHPP的SRM的概率性质。估计方法大致可分为贝叶斯估计方法和非贝叶斯估计方法。Kuo和Yang(1996)基于著名的马尔可夫链蒙特卡罗(MCMC)方法提出了一种通用的贝叶斯估计框架，用于具有和不具有有界均值函数的具有代表性的基于NHPP的SRM。Basu和
The quantitative assessment of software reliability is one of the main issues in software engineering. Over the last four decades, researchers have investigated a variety of quantitative methods for supporting decision making in real software development processes. A range of software reliability models (SRMs) have been proposed [Lyu (1996); Musa et al.(1987); Pham (2000)], taking different approaches to estimate quantitative software reliability, defined as the probability that a software system will not fail during a specified period of operation. Among these, non-homogeneous Poisson process (NHPP)-based SRMs have become widely-used in the software testing phase to assess the software reliability, identify the number of remaining faults in software, and determine the software release schedule. The mean value function is a unique parameter that supports the probabilistic property of NHPP-based SRMs. The estimation methods can be broadly classified into Bayesian and non-Bayesian. Kuo and Yang (1996) proposed a general Bayesian estimation framework based on the well-known Markov chain Monte Carlo (MCMC) method, for representative NHPP-based SRMs with and without bounded mean value functions. Basu and