Local sensitive hash-based adaptive random test method

Abstract

The invention discloses a local sensitive hash-based adaptive random test method. The method comprises the steps of 1, determining an input domain range; 2, randomly generating a first test case in the input domain and executing the first test case; 3, randomly generating k candidate test cases to form a candidate test case set; 4, selecting a next test case by adopting FSCS-ART, executing the next test case and inserting the next test case into the hash tree; and 5, repeatedly executing the steps 3 and 4 until m executed test cases are obtained or errors are found. 6, randomly generating k candidate test cases to form a candidate test case set; and 7, selecting a next test case by adopting an improved locality sensitive hash algorithm, executing and inserting the next test case into the hash tree. And repeating the steps 6 and 7 until a program error is found. Compared with the existing FSCS-ART, the LSH-ART provided by the invention has the advantages that in performance, the LSH-ARTis slightly similar to the FSCS-ART in low dimension, but the LSH-ART is better than the FSCS-ART in high dimension; in terms of time expenditure, the LSH-ART is less than the FSCS-ART in both low dimension and high dimension, and particularly, the time expenditure of the LSH-ART is even less than 10% of the time expenditure of the FSCS-ART under low failure rate.

Description

technical field [0001] The invention proposes an adaptive random testing method based on local sensitive hashing, which is used to improve the high-dimensional performance of the traditional FSCS-ART technology and reduce the time overhead of the traditional FSCS-ART technology, belonging to the technical field of software test automation. Background technique [0002] With the continuous expansion of the software market, the scale of software is increasing day by day, and the software functions are becoming more and more complex. How to ensure the quality of software products has become an important research focus, and software testing has undoubtedly become an important link to ensure software quality. For software testing, researchers have proposed many software testing techniques, among which random testing has received more and more attention because of its simple concept and automation. [0003] Random testing can detect errors that people can't expect (and this kind o...

AI Technical Summary

Problems solved by technology

When only one LSH function is generated for the input domain for calculation, if the value of w is too small, the result obtained by LSH may be very inaccurate, and even the result cannot be obtained; if the value of w is large, it may only be divided As a bucket, the algorithm is reduced to the traditional FACS_ART, which reduces the efficiency of the algorithm; in addition, even if w is properly obtained at the beginning, when there are many executed test cases, the number of executed test cases in each bucket will also be large, making Algorithmic efficiency cannot be improved enough

Images