Adaptive random testing through dynamic partitioning by localization with restriction and enlarged input domain

Abstract

Despite the importance of random testing approach, it is not used itself, but plays a core role in many testing methods. Based on the intuition that evenly distributed test cases have more chance for revealing non-point pattern failure regions, various Adaptive Random Testing (ART) methods have been proposed. A large portion of this methods such as ART with random partitioning by localization have edge preference problem. This problem would decrease the performance of these methods. In this article the enlarged input domain approach is used for decreasing the edge preference in ART through dynamic partitioning by localization with Restriction. Simulations have shown that the failure detection capability of ART through dynamic partitioning by localization with Restriction and enlarged input domain is comparable and usually better than that of other adaptive random testing approaches. © 2013 Springer-Verlag.