| Summary: | 博士 === 國立臺灣科技大學 === 電機工程技術研究所 === 86 === Generally, the reliability evaluations of fault-tolerant computer system based on the Markov model have focused on the effect of permanent fault, intermittent fault, or transient fault, separately. However, the system fault may include the effects of permanent fault, transient fault, and intermittent fault simultaneously. In this dissertation, we propose a generalized Markov reliabil-ity model, which includes the effects of permanent fault, intermittent fault, and transient fault for reliability evaluations. We also provide a neural net-work and an improved training algorithm to evaluate the reliability of the fault-tolerant systems. The desired system reliability under design is fed into theneural network and when the neural network converges, the design parameters areextracted from the weights of the neural network. The simulation results show that the neuro-based reliability models can converge faster than the other methods.The system state equations for the Markov model are a set of the first-order linear differential equations. Usually, the system reliabiliy can be evaluated from the combination of state solutions. This technique is very complicated and very difficult in the complex fault-tolerant systems. In this dissertation, we present several Grey Models(GM(1,1), DF-GM(1,1) and ERC-GM(1,1)),which offer an one-equation solution, to evaluate the reliability of computer system. It can obtain the system reliability directly and easily. But the datanumber for grey model that gets minimal error is different in each time step.Therefore, a neural network is designed on the basis of the more accuracy prediction for the grey modeling to evaluate the reliability. Finally, the simulation results show that this technique is better than the GMs in accuracy.
|
|---|
