Testing coverage based software reliability assessment incorporating effort expenditure and error generation

Due to the complexity of software systems, the testing team may be unable to completely eliminate a problem after observing a failure and the participation of numerous external elements in their development, and another fault may replace the identified fault, also known as error generation. Software developers and customers can benefit from test coverage as a software statistic since it can assist them improve the functionality of tested software and identify what additional effort is necessary to increase the reliability of the software. Software reliability growth models (SRGMs) have been cited as one of the effective methods for quantitative evaluation of software quality. The concept of effort spent and error generation are integrated into a model for testing coverage-based software reliability evaluation. Testing effort spent is supposed to follow the Weibull distribution, whereas testing coverage is described by exponential, delayed S-shaped, and logistic functions respectively. Additionally, we look into the cost requirement-based software release time for exponential functions with a reliability constraint. We introduce the genetic algorithm, which is a powerful tool for dealing with search and optimization issues. Two real failure datasets have been used to test different goodness of fit criteria for the model and their performance is evaluated using four goodness-of-fit metrics, including coefficient of determination (R ²) , mean square error (MSE), predictive power (PP) and predictive ratio risk (PRR). The acquired results outperform the perfect debugging model and demonstrate notable advancements that are fairly encouraging.