An Adaptive Random Test Method based on Variable Probability Density Function with Particle Swarm Optimization

Adaptive Random Testing (ART) is proposed to enhance the effectiveness of Random Testing (RT) based on the notation that evenly distributing test cases across the whole input domain. Adaptive Random Testing Through Test Profiles (ART-TP) has been considered as one of the most effective ART methods. Generally, the selection of probabilistic function matters significantly in terms of testing effectiveness. In this paper, to achieve better “evenly distributed”, we analyze the effect of concave-convex functions and design a new probabilistic function. Moreover, we take advantage of the particle swarm optimization (PSO) algorithm to advise test case generation and propose a new approach namely Probability Adaptive Random Testing by Particle Swarm Optimization (PART-PSO), so that the the diversity of test cases could be greatly enhanced, thus a better failure detection capability.