Integrated Evaluation Model of Software Reliability Based on Multi-information
Zhao Shou-wei, Lu Zhizhou, Sasa Ma, Liu Dongqin · 2009
The software reliability prediction model based on software quality measurement (program characteristics) and the software reliability growth model based on statistics characteristic (failure data) are the main quantification evaluation and prediction methods of software reliability. The modern creditability theory based on Bayesian estimation was introduced into the multi-information integrated modeling of software reliability evaluation and prediction. By this means, the prediction result of software reliability based on the metrics of software quality and the evaluation result of software reliability growth models based on failure data are combined reasonably via the credibility factor. Thus the extensive information related to software reliability was utilized and integrated so as to make the evaluated result of software reliability more reasonable. According to the modeling means of variance decomposition and optimal linear non-homogeneous credibility estimator, the mathematic expression of this integrated evaluation model of software reliability based on Bayesian estimation was described in detail, and the solution of credibility factor was illustrated. The data experiment testified its rationality and feasibility of this multi-information integrated model.