原文传递 Final Report: Studies in Structural, Stochastic and Statistical Reliability for Communication Networks and Engineered Systems.
题名: Final Report: Studies in Structural, Stochastic and Statistical Reliability for Communication Networks and Engineered Systems.
作者: Samaniego, F. J.
关键词: Reliability, Networks, Communication networks, Repair, Statistics, Order statistics, System signatures, Optimality, Nonparametric estimation, Asymptotics, Spo(stochastic precedence ordering), Sto(stochastic ordering)
摘要: The present report describes the research completed by Principal Investigator F. J. Samaniego under the support of ARO grant W911NF-11-1-0428 from 9/16/11 to 9/15/16. Research results in five topical areas within the field of Engineering Reliability are summarized. Contributions include: (1) new results are obtained on the treatment of repairable systems in which the theory of system signatures is extended to minimally repaired systems and new results are obtained for comparing pairs of used systems, (2) conditions are identified under which systems of different sizes in i.i.d components will have the same stochastic performance; such a result is essential in comparing the performance of two coherent systems of arbitrary sizes, (3) the notion of survival signature is applied to the comparison of the performance of systems with heterogeneous components, thereby relaxing the i.i.d assumption made on the components of coherent systems that is required for comparisons based on system signatures, (4) in comparisons among communication networks of a given size (with v vertices and n edges), new results are obtained showing that the stochastic precedence ordering (spo) can provide a total ordering of these networks with respect to the probability of total connectivity in situations where the traditional stochastic ordering (sto) fails to do so, and (5) three distinct versions of the problem of estimating component reliability from system failure-time data are treated, each resulting inconsistent estimators with asymptotically normal distributions.
报告类型: 科技报告
检索历史
应用推荐