中文摘要：

针对当前软件可靠性模型（SRM）领域失效相关性研究中存在的问题，提出了一个结合维护策略和失效相关性的SRM．首先根据维护策略的特点和2次运行之间的6种状态转换，采用了包含成功运行、失败运行和版本更新3类点的马尔可夫更新过程，在离散时间上和连续时间上分别建立了SRM；然后在此基础上，将分层思想引入模型，即在每层上对相互依赖的软件运行建模，在层与层之间对维护策略建模；最后应用上述模型得到失效间隔时间分布函数、可靠度等重要的可靠性指标．经过模型比较与分析，验证了该模型与其他SRM的内在关联是一个更具柔韧性的SRM框架．

英文摘要：

A new software reliability model （SRM） which combines maintenance policy and failure correlation is proposed to resolve the problem of current SRM. According to the characteristic of maintenance policy and six mode of state conversion, Markov renewal process of three class points including successful run and failed run as well as version update is used to establish the SRM respectively in discrete time and in continuous time. On this basis, a layer idea is introduced into the model： modeling dependent successive software runs at each layer and maintenance policy between two layers. Finally , some important reliability measurements including the distribution function of the time to failure, software reliability etc are concluded by applying this model. Through model comparison and analysis, the intrinsic consistency between the proposed model and other SRMs is validated, and the model itself can be also considered as a more flexible SRM framework.