Evolutionary generation method of test cases based on dbn

Abstract

The invention discloses a DBN-based test case generation method, and belongs to the field of software tests. The method includes: constructing different test cases through a software demand document to use the same as a training set used for training a DBN-based test case classifier; and combining a self-adaptive genetic algorithm to generate and evolve test cases. A mutation rate and a crossoverrate in genetic operations are adjusted in a self-adaptive manner through classification results of the test case classifier. Manners of selection, crossover and mutation are employed to generate newindividuals, thus the new individuals can be better generated, local extremums can be avoided, and global optimal solutions can be obtained by searching. The generated new individuals continue to be classified, test cases are output if test demand is met, and otherwise, the genetic operations continue to be carried out until the number of required test cases reaches a specified number. According to the method, the premature convergence problem existing in a process of generating test cases by applying traditional genetic algorithms can be solved, diversity of a population can be increased through a form of population classification, and efficiency of generating the test cases can be improved.

Description

technical field [0001] The invention belongs to the field of software testing, and in particular relates to a DBN-based test case evolution generation method. Background technique [0002] Software testing, which describes a process used to facilitate the identification of software correctness, integrity, safety and quality; is a review or comparison process between the actual output and the expected output. Software testing is an important link in the process of software development, and it is also a link with high costs. Statistics show that this link generally accounts for more than 50% of the total cost of software development. In recent years, software testing has been paid more and more attention and researched extensively, and the automatic generation of test data has always been the core of software testing. [0003] In order to achieve full coverage of program structure elements with a small number of test cases, some meta-heuristic search algorithms are usually us...

AI Technical Summary

Problems solved by technology

[0007] The purpose of the present invention is to improve the method of generating test cases by traditional genetic algorithm. On the basis of traditional genetic algorithm, DBN in machine learning is integrated into it, and the mutation rate and crossover rate in genetic algorithm are adaptively adjusted to solve the problem of traditional genetic algorithm. Premature Convergence Problem Existing in the Algorithm

Images