Design mode reliability analysis method based on probability model inspection

Abstract

The invention provides a design mode reliability analysis method based on probabilistic model test. The method comprises the following steps: determining the number of levels of external impact of a system and influence judgment conditions of impact levels on the system; establishing a system mixed extreme operation impact model; performing formalized description on the mixed extreme operation impact model, and establishing an impact layer continuous time Markov model; performing formalized description on different design mode abstracts, and establishing a design mode layer continuous time Markov model; establishing a corresponding relationship between system states and state transition rates of the impact layer and the design mode layer, and constructing an overall continuous time Markovmodel; defining a continuous random logic attribute specification formula of the system; and inputting the overall continuous Markov model and the continuous random logic attribute specification intothe PRISM, and carrying out reliability verification on the design mode. Reliability quantitative analysis and evaluation can be carried out on different design modes considering the impact effect, and researchers are assisted to carry out design mode level improvement and final decision making.

Description

technical field [0001] The invention relates to the technical field of software reliability verification, in particular to a design pattern reliability analysis method based on probability model inspection. Background technique [0002] In order to ensure the reliability of the system, many design patterns have been proposed, such as reconfiguration, triple redundancy and error correction coding techniques and their combination. Although there are various types of design patterns, their purpose is to improve the reliability of the system through a certain design. For example, the main idea of ​​triple-mode redundancy is to perform exactly the same task through three identical logic modules, and use a voter to switch between them. When one of the design modules fails, the system can work normally as long as the other modules continue to operate, thus avoiding the serious impact of single point failure on the system. Because the design pattern is relatively abstract, it is di...

AI Technical Summary

Problems solved by technology

However, because the design pattern is relatively abstract, it is difficult to carry out accurate modeling description. If there is not enough tool support, the traditional Markov analysis method will become difficult, especially when analyzing complex system properties.

Moreover, different design patterns have different impacts on system reliability under different objects, parameter settings and constraints. Especially, when the system is in a changing external shock environment, the law of design patterns will become more complex

From an engineering point of view, purely traditional Markov analysis methods are difficult to generalize

Images