The Research Of Adaptive Random Testing

Adaptive random testing (ART) is an enhancement to random testing for situateions where failure-causing inputs are clustered together. It has been shown by simulations and empirical analysis that ART frequently outperforms random testing. But ART's effectiveness will be greatly cut down as the dimensions of the input domain increases. Moreover, most of ART algorithms are focus on the numberic testing environment, there isn't any effective nonnumeric ART algorithm and testing tool which is based on ART.For the ART's high dimensions issue, this paper steps into the ART algorithms and focuses on the FSCS and IP. Base on this, we point out the key factors which would impact ART algorithms' effectiveness, and propose an Incremental Dimensions Independent FSCS (IDI-FSCS) algorithm. It has been proven that IDI-FSCS can highly improve the effectiveness of ART in multi-dimensions input domain. Furthermore IDI-FSCS is more efficient than FSCS algorithm.For facility, this paper proposes a grid-based measurement, which can be used to estimate ART algorithms' effectiveness quickly.In order to make ART more utility, from the Object-Oriented point of view, we define how to calculate the distance between two test cases, no matter the test cases are numberic or nonnumeric. This makes it prossible to apply ART to testing tool.According to the results listed above, we build an ART tool named Java ART Tool (JAT). This tool is more effective than random testing tool T2 and is able to test many kinds of programs which are written in Java.