Component software reliability evaluation method based on migration paths and improved Markov chain

Abstract

The invention discloses a component software reliability evaluation method based on migration paths and an improved Markov chain. The method comprises the following steps: (1) obtaining a component migration probability matrix by analyzing a software structure, drawing a dynamic migration diagram of each component, analyzing the damage degree of each component and finding out all existing migration paths; (2) testing software by using test cases of which the quality is not smaller than the lowest quantity, and counting executed times, ineffective times and ineffectiveness restored times of each component to obtain the execution probability of each component; (3) calculating according to the migration probability among the components and the execution probability of each component to obtain the occupancy rate of each component; (4) calculating the failure rate of each component; (5) calculating according to the damage degree of each component to obtain the importance degree of each component; (6) calculating according to a state migration probability matrix P to obtain the occurrence probability of each migration path; (7) evaluating the reliability of each component in each migration path; (8) evaluating the reliability of each migration path; (9) evaluating the reliability of the software.

Description

technical field [0001] The invention belongs to the field of software engineering and relates to a software evaluation method, in particular to a component software reliability evaluation method based on a migration path and an improved Markov chain. Background technique [0002] With the continuous expansion of software scale and continuous updating of functions in large-scale real-time software systems, the software structure is becoming more and more complex, and the requirements for its reliability are also getting higher and higher. Therefore, it is necessary to establish a software reliability evaluation model to evaluate the real-time software reliability. sex evaluation. The current large-scale real-time software system is composed of many software components, and the use frequency and criticality of each component are different. Therefore, it is necessary to evaluate software reliability from the perspective of software internal architecture and software components....

AI Technical Summary

Problems solved by technology

[0003] The main path-based software reliability models mainly include the following types: Shooman model, which assumes that the execution frequency and failure probability of components are known, and component failures are independent of each other, and the software reliability is calculated by calculating the total failure rate, but the The model assumptions are too strict; the Krishnamurthy and Mathur model, which calculates the path reliability evaluation of the sequence of components executed in each test run, and then averages them to obtain the system reliability evaluation, but the model cannot adapt to the The case of multiple loop body components; the Yacoub, Cukic and Ammar model uses a tree traversal algorithm to traverse all branches of the software architecture diagram to form a path, but the end condition of the model is not easy to determine

Some researchers also proposed to use the non-homogeneous Poisson model for component failure density calculation, combined with the path-based method to evaluate the reliability of software, but in this model, the execution time of components in each path cannot be counted.

[0004] The existing path-based software reliability modeling method does not consider and does not use the Markov chain (hereinafter referred to as Markov chain) model to analyze component reliability, and secondly, it does not highlight the impact of software architecture and component reliability on software system reliability. Influenced by reliability, the execution time of each component in each path has not been theoretically estimated, and it cannot be analyzed comprehensively and accurately, which directly affects the results of software reliability analysis

Images