Annighöfer, BjörnBjörnAnnighöferNil, CelenCelenNilSebald, JohannesJohannesSebaldThielecke, FrankFrankThielecke2020-08-172020-08-172015-10-28AIAA/IEEE Digital Avionics Systems Conference: 7311446 (2015-10-28)http://hdl.handle.net/11420/7063The current European space launcher ARIANE 5 has an avionics system designed in the 90's. Possible benefits of introducing Integrated Modular Avionics (IMA) to it have been investigated with a model-based optimization approach. From this investigation two novel features for avionics architecture optimization are presented, first, a method for logically structured allocation and, second, a method symmetric function allocation. IMA are state-of-the art in large aircraft. The high degree of freedom in software allocation sizing and placement makes an optimal manual planning, however, hardly possible. Optimization algorithms achieve objectively better results, but their results often lack non-functional requirements which are inherently incorporated in manual designs. For an existing avionics architecture optimization framework extensions are presented, which improve the convenience, quality, and subjective attractiveness of automatically derived architectures. All features are demonstrated for an IMA-based avionics architecture optimization for the ARIANE 5 and future launchers. The solution quality, optimization time, and subjective impressions are analyzed. It is shown how the extended optimization methods help to improve fragmentation and sensor distances by 14% and 24% respectively without degrading the primary design objective.enConference Paper10.1109/DASC.2015.7311446Other