基于失效数据统计方式的软件可靠性增长模型框架

为了更真实地描述软件失效过程，针对软件故障一旦检测出就立即排除掉这一不符合真实情况的假设条件，考虑到检错过程和排错过程的异步性，提出了计入重复故障的可靠性增长模型框架（RFM）和不计重复故障的可靠性增长模型框架（NRFM），并指出了每个框架的适用范围．最后以NRFM为例，给出一个新的可靠性增长模型．通过公开发表的一组数据集上的拟合实验和预测实验表明，与其他模型相比，该模型具有很好的拟合能力和预测能力．

SRGM Framework Based on Failure Data Statistical Pattern

For purpose of describing the software failure process more really, directing at the false hypothesis that the faults would be immediately removed, once they are detected, and taking the asynchronism between checking and Debugging processes into consideration, this paper proposes a reliable growth model framework in which the repetitive faults are reckoned in （RFM） and the one where the repetitive faults are not reckoned in （NRFM）, and defines the applicable range for each framework as well. Finally, by taking the NRFM as an example, a novel reliable growth model is given in this paper. By performing the fitting experiment and prediction experiment in terms of a set of DS published, it is shown that this proposed model is of excellent capability of fitting and prediction, compared with other models.