Evolutionary Testing Research at Berner & Mattner

For processes that are controlled by embedded systems, more and more software is needed in vehicles that has to be put through its paces in order to ensure safety and reliability. In a modern car, this software often contains more than ten million lines of software code. The difficulty is to define meaningful tests to eliminate as many potential failure causes as possible already in the software test phase. Berner & Mattner was already involved in the fundamental research within the European EvoTest project and is now working together with UCL CREST to further develop the technology for industrial applications. A predominant goal of evolutionary testing is to automatically select the most demanding test scenarios.



Professor Mark Harman, director of the UCL CREST, comments: Using Darwin's theory of evolution, we can generate exceptionally demanding test cases. We create a virtual world in which car testing scenarios have to compete with one another. In this world of test scenarios, a test case has to be demanding in order to survive. This survival of the fittest has the effect of generating especially interesting test cases to be used in testing the automotive software.



Dr. Joachim Wegener of Berner & Mattner explains the economic importance of the project funded by the Engineering and Physical Sciences Research Council (EPSRC) in the UK: Testing is one of the most costly activities within the development of embedded systems. Search-based testing techniques have the potential to fully automate the testing of embedded systems. This will allow for significant cost savings and increased product quality. Our cooperation with UCL is aimed at further developing these technologies for a regular usage in industrial practice.