An Approach on Automatic Test Data Generation with Predicate Constraint Solving Technique

Predicate constraint solving technique is an important method of automatic test data generation. By analyzing the properties and disadvantages of predicate constraint solving technique, three theorems are proposed and proved. Based on them, a new approach on automatic test data generation is presented. The linear predicate on a given path is used directly to construct the linear constraint system. Only the linear arithmetic representation for nonlinear predicate function is required to compute. Theoretical analysis and practical testing show: for the linear path, the initial input and the iteration are not required. The path is feasible if the linear constraint system can be solved, and the values of the input variables obtained are the desired test data. Otherwise, the path is infeasible. Considering the nonlinear path, test data can also be obtained by a number of iterations, or it can be guaranteed that the path is infeasible to a great extent.