An efficient reduction approach to test suite

The paper presents a novel reduction approach to test suite in regression testing. Our approach selects a test case from those redundant test cases with the same rank according to the boundary coverage taken as second coverage criterion. Different from the previous researches for the selection of the second coverage criterion to assist test suite reduction, our approach bases on our empirical point of view that some faults in the program lie on its boundary of the program. Therefore test cases picked by our approach have the higher probability to find software errors. An algorithm is proposed to realize our approach. In addition, we also discuss the relationship between our approach with three traditional heuristics. Through a simple exampke, we illustrate our algorithm for test suite reduction.