State based testing using swarm intelligence

Automatic data flow testing scrutinizes the flow of data within models by using data flow analysis rules. To certify accurate data flow within states we have to contemplate the data values. The investigation of data flow forms a foundation of data flow testing by bearing in mind defines and uses of the variables. Empirical studies have shown that existing state- based approaches are not competent in uncovering state based faults. State-based testing scrutinizes state changes and its behavior without concentrating on the internal details, thus data faults remained uncovered. It has been observed that many state based approaches don´t offer complete definition-use path complete coverage and also are ineffective in terms of detection of data flow faults. Our work starts from these observations to view automatic data flow analysis problem to solve with heuristic technique. In this research, an approach is presented for automatic testing of data flow of UML state machine models and automatically generates test cases. Data flow testing problem is view as an optimization problem while selecting optimal number of feasible test cases in providing complete or maximum def-use paths coverage. An optimal solution is investigated by exploiting the heuristic search technique Ant Colony Optimization Algorithm. We implemented this approach in a tool named data flow testing tool. Effectiveness of our proposed approach is analyzed by applying it to UML state based models representing the dynamic system behavior. Experimentation is performed for the validation of this approach.