Applying Feedback Information for Random Partition Testing

Random testing is a basic strategy in software testing. Partition testing is another basic testing strategy that separates input domain into disjoint sub domains. However, the effectiveness of random testing and partition testing should be improved by using feedback information. Feedback-controlled testing has been introduced in adaptive testing strategy. Computational overhead become a major issue in adaptive testing strategy. The algorithms for parameter estimation have high computational overhead. In this paper, a simplified parameter estimation is introduced by applying feedback mechanism. Experimental results show that the studied method can improve the cost-effectiveness of random testing with lower computational overhead than the adaptive testing strategy.