A Test Prioritization Algorithm That Cares for "Don't Care" Values and Higher Order Combinatorial Coverage

Abstract

The efficiency of prioritization algorithms depends on how early the faults are detected. In this paper, we present a novel prioritization algorithm for combinatorial testing. Our approach takes ACTS tool generated test cases with "don't care" values as the starting point and refines them for increased effectiveness without increasing the number of test cases. Our algorithm maximizes the number of higher order combinations tested, by filling the "don't care" values in the test suite effectively. It also orders the test cases using a cost function that includes higher order coverage, thereby achieving early fault detection. The effectiveness of our algorithms is demonstrated by performing a comparative evaluation using the metric t way Rate of Fault Detection, on 2 real life case studies and numerous synthetic covering arrays of different sizes. The results show that our algorithms perform better in terms of covering higher order pairs and also faster.