The common assumption for most existing\nsoftware reliability growth models is that fault is independent\nand can be removed perfectly upon detection. However,\nit is often not true due to various factors including software\ncomplexity, programmer proficiency, organization hierarchy,\netc. In this paper, we develop a software reliability model\nwith considerations of fault-dependent detection, imperfect\nfault removal and the maximum number of faults software.\nThe genetic algorithm (GA) method is applied to estimate\nthe model parameters. Four goodness-of-fit criteria, such as\nmean-squared error, predictive-ratio risk, predictive power,\nand Akaike information criterion, are used to compare the\nproposed model and several existing software reliability\nmodels. Three datasets collected in industries are used to\ndemonstrate the better fit of the proposed model than other\nexisting software reliability models based on the studied criteria.
Loading....