Based on the Random Vector Mirror Method Improve the ART Algorithm
Hou Shaofan1,2), Yu Lei1,2), Li Zhibo1,2), and Zhang Xinglong1,2)
1) (PLA Information Engineering University, Zhengzhou 450001) 2) (State Key Laboratory of Mathematical Engineering and Advanced Computing, Zhengzhou 450001)
To cope with problem that the low randomness of test cases, generated by the mirror function of existing mirroring adaptive random testing (MART), dynamic mirror adaptive random testing (DMART) algorithms and so on, make the effectiveness of the algorithms so declined obviously in varying degrees. A mirror method based on random vector is proposed to improve the adaptive random testing (ART) algorithms. Firstly, the traditional mirror function is improved by introducing the random vector to enlarge the diversity between the mirror test cases. And then, the random vector mirror function is applied to the mirror method to improve the ART algo-rithms. The experimental results show that, use the random vector mirror method can improve the effectiveness of mirror algorithms visibly, and this algorithm enhances prominently in comparison with the original algorithm in efficiency.
software testing; random testing; adaptive random testing; mirror; mapping; divide and conquer; ramdom vector