VPPS-ART: An Efficient Implementation of Fixed-Size-Candidate-Set Adaptive Random Testing using Vantage Point Partitioning Strategy

As an enhanced version of Random Testing (RT), Adaptive Random Testing (ART) aims to improve the failure detection effectiveness of RT by distributing the test cases more evenly in the input domain. Many ART algorithms have been proposed based on different criteria. Among them, the Fixed-Size-Candidate-Set ART (FSCS-ART) is one of the most effective and classical algorithms. FSCS-ART ensures high failure detection effectiveness using selecting the candidate test case which farthest from the previously executed test cases as the next test case. Although FSCS-ART has good failure detection effectiveness, it also has some drawbacks, such as computational overhead problem. In this paper, we propose an enhanced version of FSCS-ART, namely Vantage Point Partitioning Strategy based ART (VPPS-ART). VPPS-ART attempts to solve the computational overhead problem of FSCS-ART using vantage point partitioning strategy and ensures the failure detection effectiveness of FSCS-ART. VPPS-ART achieves the partitioning of the input domain space by using a Vantage Point tree (VP-tree) and finds the nearest executed test cases of a candidate test case in the partitioned sub-domains, which reduces the time overhead significantly compared to the entire input domain search computation. Besides, to match the dynamic insertion process of FSCS-ART, we modify the structure of the traditional VP-tree to support dynamic data. The simulation results present that VPPS-ART has a great lower time overhead compared to FSCS-ART, and also guarantees similar or better failure detection effectiveness than FSCS-ART. The VPPS-ART also shows strength in comparison with the KDFC-ART algorithms, a series of enhanced ART algorithms based on KD-tree. Our empirical studies also reveal that VPPS-ART is more cost-effective compared to FSCS-ART and KDFC-ART.