多阶段多状态系统可靠性的动态故障树模块化分析方法

项目名称： 多阶段多状态系统可靠性的动态故障树模块化分析方法

项目编号： No.61272130

项目类型： 面上项目

立项/批准年度： 2013

项目学科： 自动化技术、计算机技术

项目作者： 莫毓昌

作者单位： 浙江师范大学

项目金额： 82万元

中文摘要： 多阶段多状态系统（MPMSS）是一类具有阶段性运行特征和多状态部件的复杂计算系统，其可靠性分析受到学术界高度重视，但已有分析方法在效率和精度方面不能很好的满足实际需求。本项目针对已有研究在模块化方法、动态模块分析、模块联合概率求解和模块化故障树决策图构造等方面存在的问题，重点研究：多粒度模块划分和特征依赖模块化方法选择、各种粒度模块化下通用模块联合概率求解、基于补充变量和MRGP 的高效动态模块分析、基于分治Ebendt 搜索和启发式重组的模块化故障树决策图构造。项目预期将获得高效的MPMSS 可靠性模块化混合分析方法，从而扩大可分析MPMSS 的规模和复杂度、提升可靠性分析的准确性和效率。项目的研究成果对于完善复杂计算系统可靠性评估的理论研究具有重要意义，可以为关键应用领域中实时容错计算系统的可靠性评估提供切实可行的解决方案，为冗余系统结构和容错机制等高可靠性设计决策的制定提供科学依据。

中文关键词： 多阶段多状态系统；动态故障树；模块化；决策图；

英文摘要： Multi-phase multi-state systems (MPMSS) are complex computing systems in which both the systems and/or their components may be subject to multiple, consecutive, non-overlapping phases of mission execution and may exhibit multiple performance levels or states. The reliability analysis of MPMSS has been considered tasks of primary relevance that has been widely investigated and far from being solved. This project tries to cope with the problems has been found in modular solution, dynamic module analysis, module joint probability calculation and decision diagram generation of modularized fault tree. In this project we propose several modular solutions with different granularities, structure characteristics- dependent selection of modular solutions, generalized module joint probability calculation, dynamic module analysis based on supplementary variable method and Markov regenerative process method, and decision diagram generation with divide-conquer-based Ebendt search or shuffling heuristics. The obtained hybrid modular solution can be used to analyze larger and more complex real-world MPMSS instances, and can significantly improve both the accuracy and efficiency of MPMSS reliability analysis. This research work is important for the academic research on reliability analysis of complex computer systems. The resear

英文关键词： Multi-Phase Multi-State Systems；Dynamic Fault Tree；Modularization；Decision Diagram；