CSR-EHCS(EHS), SM: Efficient Dynamic Combinatorial Models (EDCM) for Reliability Analysis of Complex Dynamic Systems

在系统可靠性分析领域，动态和相关的行为，如功能依赖，级联故障，状态依赖已被公认为是一个显着的贡献，在整个系统的可靠性问题。然而，与动态和相关的行为，所产生的动态系统可靠性模型不能有效地和准确地解决现有的基于状态空间的方法，如马尔可夫方法。 这是由于两个关键的限制：状态空间爆炸问题，以及无法处理任意故障分布。该项目探索有效的组合模型，形式化方法和转换方法，以解决这两个限制，提供一个可行的解决方案，以实际的大型计算机为基础的系统的精确可靠性分析具有复杂的动态行为。本研究包含四个主要部分：1）动态和相关行为的可靠性建模和评估，2）所提出的可靠性模型和评估方法的验证和性能分析，3）案例研究/工业应用，以及4）基于GUI的可靠性软件工具开发。通过该项目开发的新的可靠性模型和评估方法是对计算机系统可靠性知识体系的基本贡献。该项目通过其对计算机系统可靠性研究，研究生和本科生教育以及复杂和动态工业系统的可靠设计的贡献产生了更广泛的影响。通过一个项目网站，可公开获取项目产生的出版物、软件工具和与工具有关的材料。

In the area of system reliability analysis, dynamic and dependent behavior such as functional dependence, cascading failures, and state dependence has been recognized as a significant contribution to problems in overall system reliability. However, with the incorporation of the dynamic and dependent behavior, resulting dynamic system reliability models cannot be efficiently and accurately solved by existing state space based methods such as Markov methods. This is due to two key limitations: the state-space explosion problem, and the inability to handle arbitrary failure distributions. This project explores efficient combinatorial models, formal methods, and transform methods to address these two limitations, providing a viable solution to the accurate reliability analysis of practical large-scale computer-based systems with complex dynamic behavior. The research has four major components: 1) reliability modeling and evaluation of dynamic and dependent behavior, 2) verification and performance analysis of the proposed reliability models and evaluation methods, 3) case studies/industrial applications, and 4) GUI-based reliability software tool development. The new reliability models and evaluation methods developed through this project are fundamental contributions to the body of knowledge on the computer system reliability. This project has broader impact through its contributions to computer system reliability research, graduate and undergraduate education, and the reliable design of complex and dynamic industrial systems. A project website provides publicly available access to the publications generated from the project, software tools and tools-related materials.