An Integrated Framework for Developing Discrete‐Time Modelling in Software Reliability Engineering

Abstract

In the software reliability engineering literature, few attempts have been made to study the fault debugging environment using discrete-time modelling. Most endeavours assume that a detected fault to have been either immediately removed or is perfectly debugged. Such discrete-time models may be used for any debugging environment and may be termed blackbox, which are used without having prior knowledge about the nature of the fault being debugged. However, if one has to develop a white-box model, one needs to be cognizant of the debugging environment. During debugging, there are numerous factors that affect the debugging process. These factors may include the internal, for example, fault density, and fault debugging complexity and the external that originates in the debugging environment itself, such as the skills of the debugging team and the debugging effort expenditures. Hence, the debugging environment fault removal may take a longer time after having been detected. Therefore, it is imperative to clearly understand the testing and debugging environment and, hence, the urgency to develop a model. The model ought to take into account the fault debugging complexity and incorporate the learning phenomenon of the debugger under imperfect debugging environment. This objective dictates developing a framework through an integrated modelling approach based on nonhomogenous Poisson process that incorporates these realistic factors during the fault debugging process. Actual software reliability data have been used to demonstrate applicability of the proposed integrated framework.

Software warranty cost optimization under imperfect debugging

“…Recently, some of the researchers have incorporated error generation into their models (Aggarwal et al , 2019; Gandhi et al , 2019; Gautam et al , 2018; Jain et al , 2012; Kapur et al , 2011a; Li and Pham, 2017; Pham, 1996). The parameter estimation results (Table II) are consistent with the results presented in the literature (Johnston et al , 2019; Kapur et al , 1994, 2011a; Kapur and Younes, 1996; Shatnawi, 2016; Tamura and Yamada, 2018; Yamada et al , 1986). The value of error generation parameter α is 0.247 and probability of perfect debugging r is 0.785.…”

“…As till now, this data set has been cited quite more than 100 times. Very well-known research studies have validated their model on the same data set (Johnston et al , 2019; Kapur, Pham, Anand and Yadav, 2011; Kapur and Younes, 1996; Ohba, 1984; Pham and Pham, 1991; Shatnawi, 2016; Tamura and Yamada, 2018). Going with their judgment, we have also selected this data set.…”

Discrete Time-Space Stochastic Mathematical Modelling for Quantitative Description of Software Imperfect Fault-Debugging with Change-Point

“…In other words, assessing reliability is pointless if the software-reliability dataset during debugging does not exhibit growth. The trend tests that are widely adopted with f fault debugged are [2], [11], [33]:…”

“…When the PRE value is positive/negative, the model overestimates or underestimates the future debugging phenomenon. Acceptable values are within the range ±10% [1], [2], [16].…”

Release Policy, Change-Point Concept, and Effort Control through Discrete-Time Imperfect Software Reliability Modelling

“…Whereas, a computer test run is a set of software input variables arranged in a certain manner to test the functional performance of a particular part of the software system. Therefore, discretetime models are more realistic and often provide better fit than their continuous-time counterparts [2][3][4][5][6][7][8][9].…”