Research For Software Behavior Modeling Based On Finite Automaton

This paper is part of the research content "863 Project: software systems analysis techniques and safety reverse system", by of research this hot issue, that is software security, found that the software behavior modeling in software analysis, testing and validation are important applications. Good software behavior model to capture more useful information more accurately, to ensure that the software analysis, testing and validation of the usefulness and effectiveness. The ultimate goal of behavior analysis is characterized by behavioral analysis of the data to obtain the behavior of the semantic description and understanding, and behavioral modeling and testing is a critical step in achieving the objectives.In this paper, modeling software behavior is based on the research, by comparing the DMC, online detection algorithm and GK-tail algorithm for modeling these types of common, I hava found that GK-tail algorithm is of higher value, and the technical focus of the algorithm is the data value constraints and the interaction between software components to form a model of higher accuracy.This paper has improved GK-tail algorithm. The extended finite automaton as a modeling tool, improved dynamic analysis tools, the use of Daikon and ESC / JAVA tools to get the combination of extended finite automaton boundary constraints, to generate models to improve the accuracy of the software analysis, testing and validation provide better assurance that the resulting behavioral model can capture more and accurate information. Finally, the design and implementation of the software system validation improve the behavioral modeling algorithm, and verify its feasibility.