基于HMM的系统可靠性建模研究

传统的系统可靠性分析需要检测系统中所有元件的故障状态，并不适用予系统的定期维护和保养检查。隐马尔可夫模型(HMM)是一种双重随机过程，能够解决随机不确定问题。通过对系统关键点的检测，经过复杂的网络运算综合得到系统状态的检测参数，给出了实现检测的相关网络模型以及相应的算法。

Research of System Reliability on Hidden Markov Model

In order to assure the system at the non - fault state, the traditional reliability method which is detected all the components of the system can be often applied. Plenty of the system information are concentrated by this method. But it has some disadvantages, such as the long period, the large cost, etc. So, this method is often applied in the regular inspecting and maintenance. But, it is not fit in the extraordinary condition. Hidden Markov Model (HMM) is a double random process. It synthesizes the character of neural network and Markov process in orde,r to simulate the change of signal's statistic character. In this paper, an available model based on Hidden Markov Model is established. The arithmetic and the structure of the system is given. In the software failure prediction, this method is exceeding to others results. Parameter Y can be believed as sensitive parameter, and the long - storage test is necessary.