Identification of Fuzzy Relations with Multiple Hyperplanes by Genetic Algorithm in Structural Design |
| |
Authors: | Yasutaka Ohmachi Hiroshi Kawamura Akinori Tani Hajime Yoneda |
| |
Affiliation: | Department of Architecture and Civil Engineering, Faculty of Engineering, Kobe University, Rokkodai, Nada, Kobe 657, Japan;Structural Engineering Division, Shimizu Corporation, Yokohama 231, Japan |
| |
Abstract: | Abstract: This paper proposes a method for formulation and quantification of human subjective evaluations and uncertain numerical values by using multiple hyperplanes based on the results of questionnaires distributed to structural designers that are composed of four questions about structural design and planning, i.e., ( Q 1) structural damage , ( Q 2) structural economy in a macroscopic sense , ( Q 3) structural safety, and ( Q 4) opening patterns in facade. In this paper, some hyperplanes in multidimensional spaces are used as a method to identify fuzzy relations (states) in intelligent fuzzy networks for seismic structural planning. A genetic algorithm is employed as a search method for optimal formulations and is proved to be very effective as a method for such an identification. |
| |
Keywords: | |
|
|