Computational grid systems are becoming popular as computational environments for dealing with large-scale applications in various areas. Conventional reliability models have some common assumptions that cannot be applied to the grid systems. This is the first study that introduces a methodology for efficient construction of Bayesian network (BN) models for estimating grid system reliability. Our proposed method uses historical data about the grid system to be modeled and constructs the BN without the need for human intervention. Unlike previous studies, our methodology does not rely on any assumptions about the grid system.
展开▼
