The present paper describes a genetic algolrithm (GA) based solution strategy for the multidisciplinary design of helicopter rotor blades. A decomposition approach is used to partition the large scale design problems into smaller, more tractable sub-problems. A model based on the parallel implementation of genetic algorithms is shown to be naturally amenable to this approach, providing increased computational efficiency, and a natural approach to account for the coupling between temporarily decoupled sub-problems. The design space for this problem is nonconvex, and the underlying analysis is inherently nonlinear and computationally demanding. The proposed procedure draws upon the use of artificial neural networks (ANN's) to identify a topology for problem decomposition, and for developing function approximations for use in optimization.
Genetic algorithmus in multidisciplinary rotor blade design
Genetische Algorithmen bei der multidisziplinären Konstruktion von Rotorflügeln
1995
11 Seiten, 7 Bilder, 6 Tabellen, 23 Quellen
Conference paper
English
Genetic Algorithms in Multidisciplinary Rotor Blade Design
British Library Conference Proceedings | 1995
|Parallel Genetic Algorithm Implementation in Multidisciplinary Rotor Blade Design
Online Contents | 1996
|