Enhanced Mirror Adaptive Random Testing Based on I/O Relation Analysis

摘要

Adaptive Random Testing (ART) is an effective improvement of Random Testing (RT). It is based on the observation that failure-causing inputs tend to be clustered together and the generic characters of typical failure patterns. By far, many ART algorithms have been developed. However, most of them have boundary effect and make use of less information of the specification. For these two issues, we propose two enhanced ART algorithms based on the idea of virtual images and the I/O relations of the program under testing respectively. Our simulation experiments show that the first algorithm can avoid the boundary effect of previous ART methods and the second can improve the failuredetection effectiveness of ART. Eventually, we obtained unexpected results in the last experiment using an integration of two new algorithms.