Software components are measured in off the shelf at a great rate to develop reliable software applications. In component based systems, reliability estimation models are developed based on reliability of individual components and their transition probabilities. In this paper, we consider Scenario based Reliability Estimation using the hierarchical model (Discrete Time Markov Chain). In hierarchical model, variance of the number of visits to each component makes the system reliability closer to the estimated reliability value. We propose a reliability estimation model for the whole application. Component Dependency Graph of the systems, Reliability of individual components and transition probabilities between two components are considered as input for the estimation. Discrete Time Markov Chain is used to generate the mean and variance of the number of visits of the individual component from the initial component. This paper illustrates the reliability prediction and sensitivity analysis with examples.

大量评估了软件组件的现成性，以开发可靠的软件应用程序。在基于组件的系统中，可靠性估计模型是根据各个组件的可靠性及其转换概率开发的。在本文中，我们考虑使用分层模型（离散时间马尔可夫链）的基于场景的可靠性估计。在分层模型中，对每个组件的访问次数的变化使系统可靠性更接近于估计的可靠性值。我们提出了整个应用的可靠性评估模型。系统的组件依赖性图，单个组件的可靠性以及两个组件之间的转换概率被视为估算的输入。离散时间马尔可夫链用于生成单个组件与初始组件之间的访问次数的均值和方差。本文通过实例说明了可靠性预测和灵敏度分析。