Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

Evolutionary testing in component-based real-time system construction

: Groß, H.-G.; Mayer, N.

Cantú-Paz, E.:
Genetic and Evolutionary Computation Conference, GECCO 2002. Late-Breaking Papers
New York, 2002
pp.207-214 : Ill., Lit.
Genetic and Evolutionary Computation Conference (GECCO) <4, 2002, New York/NY>
Conference Paper
Fraunhofer IESE ()
analysis; real-time system; genetic algorithm

Execution time analysis is an essential verification activity during real-time system development. This activity can be performed dynamically through optimisation-based analysis techniques such as evolutionary testing. Evolutionary testing is already successfully used under the traditional procedural development paradigm. This paper is an initial attempt in making evolutionary testing also applicable under the more recent object-oriented, component-based software development paradigm. Here, the application of evolutionary testing is more difficult compared with the traditional procedural development approches because component-based architectures are inherently encapsulated, and often represent state machines. This work proposes a solution that makes evolutionary testing available in component-based real-time system construction. It is based on built-in testing interfaces and on the execution of an object's invocation history through evolutionary testing.