Efficient Test Data Generation for Variables with Complex Dependencies

Abstract

This paper introduces a new method for generating test data that combines the benefits of equivalence partitioning, boundary value analysis and cause-effect analysis. It is suitable for problems involving complex linear dependencies between two or more variables. The method aims at covering all semantic dependencies plus all (n-dimensional) boundaries with a minimum set of test data. To overcome the mathematical complexity of the method, a main goal of the research project was to develop a user-friendly tool that allows users to specify dependencies in a simple language and generates appropriate test data automatically. The tool has been incorporated into the IDATG (Integrating Design and Automated Test case Generation) tool-set and validated in a number of case studies.