Modifying test suite composition to enable effective predicate-level statistical debugging

Published in NASA Formal Methods: 4th International Symposium, NFM 2012, Norfolk, VA, USA, April 3-5, 2012. Proceedings 4, 2012

In order to effectively deal with increased complexity and production pressures for the development of safety-critical systems, organizations need automated assistance in program analysis and testing. This need is intensified for systems that make heavy use of floating-point computations. Challenges related to the use of floating-point computations exist in the fields of testing, formal verification and debugging. While testing and formal verification provide mechanisms to identify possible failures within safety-critical systems, debugging techniques are employed to automatically isolate the cause of the failure. Recent advances in predicate-level statistical debugging have addressed localizing faults due to floating-point computations. Here, we present a methodology to modify the composition of a test suite to enable predicate-level statistical debuggers to more effectively isolate the causes of failures.

Recommended citation: Gore, Ross; Reynolds, Paul F. (2012). "Modifying test suite composition to enable effective predicate-level statistical debugging." NASA Formal Methods: 4th International Symposium, NFM 2012, Norfolk, VA, USA, April 3-5, 2012. Proceedings 4. 70-84.
Download Paper