White-box test case generation based on improved genetic algorithm

Abstract

Some intermittent or transient failures are particularly difficult to diagnose in highly complex and interconnected systems. This paper focuses on the use of genetic algorithms for automatically generating software test cases. In particular, this research extends a newly improved genetic algorithm, which adopts back propagation algorithm for local fine-tuning in the final link, and speeds up access to the best population. The various approaches offer opportunities for performance improvements that make these techniques more scalable for realistic applications. © 2012 Springer-Verlag GmbH.