Heuristics for Improving Model Learning Based Software Testing

Abstract

In order to reduce the cost and provide rapid development, most of the modern and complex systems are built integrating prefabricated third party components COTS. We have been investigating techniques to build formal models for black box components. The integration testing framework developed by our team leaves several open strategies; we will be investigating variations of these open strategies to enhance applicability. We are investigating the heuristics to improve the existing methodologies for learning black boxes and integration testing. We are addressing the counter-example part of the learning algorithm for improvements and are examining different techniques to identify the counterexamples in a more efficient way.