中文摘要：

软件可靠性增长模型中测试阶段和操作运行阶段环境的不同导致了两个阶段故障检测率的不同.非齐次泊松过程类软件可靠性增长模型是评价软件产品可靠性指标的有效工具.在一些非齐次泊松过程类模型中,有些学者提出了常量的环境因子,用来描述测试环境和运行环境的差别.实际上,环境因子应该是随时间变化的变量.考虑了运行阶段和测试阶段环境的不同,根据实测数据得到了变化的环境因子,并且根据测试阶段的故障检测率和变化的环境因子,转化得到了操作运行阶段的故障检测率.考虑到故障的排除效率和故障引入率,从而建立了一个既考虑运行环境和测试环境差别,又考虑故障排除效率和故障引入率的非齐次泊松过程类软件可靠性增长模型（PTEO-SRGM）.在两组失效数据上的实验分析表明,对这组失效数据,PTEO-SRGM模型比G-O模型等模型的拟合效果和预测能力更好.

英文摘要：

The testing and operation environment may be essentially different, thus the fault detection rate of testing is different from that of the operation phase. Software reliability growth models （SRGMs） based on the non-homogeneous Poisson process （NHPP） are quite successful tools that have been proposed to assess the reliability of software. The constant environmental factor is proposed by some authors to describe the mismatch between the testing environment and operation environment in SRGMs of NHPP. Actually, the environmental factor ought to be varying with testing time. The varying environmental factor with time can be derived from actual failure data set. The fault detection rate （FDR） of operation is transformed from that of testing phase and varying environmental factors, considering the fault remove efficiency and fault introduction rate, and then an NHPP model PTEO-SRGM is presented. Finally, the unknown parameters are estimated by the least-squares method based on two failure data sets. Experiments show that the goodnes-of-fit and predictive power of PTEO-SRGM is better than those of other SRGMs on these two data sets.