A genetic algorithm to devise test case sequences based on a state machine diagram and data flow information

Recently, a lots of research attention has been paid to Model-Based Testing, in which test cases are derived from the whole or a part of a model that describes static and dynamical behaviors of a system. State-Based Testing is one of the most important research topics in Model-Based Testing. The objective of this work is to research a new state-based testing approach to help software testers detect defects in implementing software systems as early as possible. To this end, the test cases are derived from a UML state machine diagram, the output of the design phase of a typical lifecycle of software development. In this work, we apply a Genetic Algorithm (GA) to select and prioritize these test cases in sequences to satisfy different user constraints such as time and coverage. In particular, the GA is expected to generate test case sequences that increase their cumulative coverage of data flow information contained in operation contracts as quickly as possible. To evaluate the usefulness of our approach, GA-generated test case sequences were compared with randomly-generated test case sequences in terms of data flow coverage and mutant effectiveness. The experimental results demonstrates that the GA-based approach is useful and effective for creating test case sequences to detect defects based on different user constraints.