Three techniques to improve standard EMOA (multi-objective optimization algorithm) on two different applications from aerospace industry are presented. These highlight the latest achievements from the EA/EMOA field, emphasizing the necessity of meaningful, statistical significant results, and maybe foster the application of such recent achievements. All presented concepts are able to provide improved results compared to the basic algorithms considered. The snapshot of only two aerodynamic applications is too small to prove a general applicability. However, the concepts are sound and valuable. To this end, these concepts are expected to be highly competitive even on other aerodynamic applications and beyond. All presented methods are highly connected to statistical modelling and testing. It can be seen as a major development within the EA/EMOA field in recent years that such issues are becoming more and more important and on the way to become state-of-the art approaches. However, recent applications from aircraft industry show an opposite tendency. In contrast to focussing on more experiments to receive meaningful, statistically significant results, the applications themselves rely on more and more complex simulations and optimization models, which require an increasing amount of computational resources. However, experts from the optimization field certainly need realistic test cases to provide developers with highly applicable, successful optimization methods. There is an implicit trade-off between having to spend increasing resources for more complex test cases versus having the chance to focus on properly planned and sufficiently repeated experiments. This trade-off is an issue for practitioners within aircraft development with no specific expertise in optimization. The hierarchical models are promising and straightforward approach, which pick up the metamodelling approach and expand it to multiple models of different complexities. To this end, Kriging as well as CFD simulations are just two levels within the model hierarchy that consists of even more models sharing different properties. Using such hierarchical models, more complex tasks can be considered.


    Access

    Access via TIB

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    Advanced concepts for multi-objective evolutionary optimization in aircraft industry


    Additional title:

    Fortgeschrittene Konzepte zur multikriteriellen, evolutionären Optimierung in der Luftfahrtindustrie


    Contributors:
    Naujoks, B. (author) / Trautmann, H. (author) / Wessing, S. (author) / Weihs, C. (author)


    Publication date :

    2011


    Size :

    16 Seiten, 9 Bilder, 7 Tabellen, 44 Quellen




    Type of media :

    Article (Journal)


    Type of material :

    Print


    Language :

    English