An enhanced zero-one optimal path set selection method

Abstract

One issue in structural program testing is how to select a mimimal set of test paths to meet certain test requirements. This is referred to as the optimal path set selection problem. The previously proposed zero-one optimal path set selection method is a generalized method that can be applied to most coverage criteria. However, the major drawback of this method is that for a large program the computation may take ten or more hours because the computation is exponentially proportional to the number of candidate paths and proportional to the number of components to be covered. To alleviate the drawback, this paper enhances the method by (1) defining five reduction rules, and (2) reusing previously selected path sets to reduce both the number of candidate paths and the number of components to be covered. Since both the number of candidate paths and the number of components to be covered are reduced, the computation time can be greatly reduced. In addition, an efficient approach to handling infeasible paths is proposed. An evaluation of the enhanced zero-one method, the original zero-one method, and a greedy method is executed and the result is presented.