An extended finite state machine test data generation method based on variable partition

Abstract

The invention discloses an extended finite state machine test data generation method based on variable division, and belongs to the technical field of software test, in particular to the technical field of model-based test. In the test based on extended finite state machine, how to improve the searching efficiency of the algorithm is a key problem in obtaining test data by heuristic method. The invention provides a method for automatically dividing independent input variables in a test path on the basis of a genetic algorithm. This method is characterized by in-depth analysis of variables andvariables in the model, The relationship between a variable and a predicate condition, determining an independent input variable that does not affect the predicate condition in the subpath and its initial partition migration, At the same time, a dynamic executable model framework is developed to provide the information feedback needed by the algorithm, and then the segmented input variables are gradually separated from the individual in the search process of genetic algorithm to achieve the automatic reduction of search space, quickly generate test data of the target test path, and improve theefficiency of test data generation.

Description

technical field [0001] The invention belongs to the technical field of software testing, and in particular relates to the technical field of model-based testing, which is used to determine and segment irrelevant input variables in the generation of extended finite state machine test data. Specifically, it refers to an extended method based on variable segmentation. Finite state machine test data generation method. Background technique [0002] Software testing plays a key role in ensuring software quality. A model is an abstract description of a software system. Model-Based Testing (MBT) uses a formal method to determine where software errors lie, thereby ensuring the correctness and reliability of the software to be tested. Because MBT is not aimed at a certain program code itself, it is widely used in product testing in the early stages of software life cycle, especially suitable for the early evaluation of safety-related systems such as aerospace software. It is one of t...

AI Technical Summary

Problems solved by technology

[0005] 1) When solving combinatorial optimization problems, the complexity of the algorithm is proportional to the size of the problem. For the generation of test data, when the number of input variables is large, the time spent on searching will increase significantly, even within a limited time Difficulty generating valid test data

[0006] 2) When using the heuristic method to solve the value of the input variable that satisfies the constraint condition, the value of all input variables will be changed at the same time to make the cost function approach the correct direction, because only one constraint condition is considered at a time, so the solution that satisfies the constraint condition After the variable value of the condition, it is necessary to search for the variable value that satisfies the next constraint condition on the basis of the current variable value. If some variables do not meet the next condition, the value of all variables will be changed to search again. This search process will lead to another Some variables that satisfy all the constraints no longer satisfy, increasing the probability of multiple backtracking

Images