Prioritized Constraints with Data Sampling Scores for Automatic Test Data Generation

Many automatic test data generation approaches use constraint solvers to find data values. One problem with this method is that it cannot generate test data when the constraints are not solvable, either because there is no solution or the constraints are too complex. We propose a constraint prioritization method using data sampling scores to generate valid test data even when a set of constraints is not solvable. Our case study illustrates the effectiveness of this method.