Computational & Technology Resources
an online resource for computational,
engineering & technology publications
Civil-Comp Proceedings
ISSN 1759-3433
CCP: 11
Edited by: B.H.V. Topping
Paper VI.1

Adaptive Connections of Multitype Structural Descriptors in a TLS for Structural Analysis

R. Fruchter and J. Gluck

Department of Civil Engineering, Technion-Israel Institute of Technology, Haifa, Israel

Full Bibliographic Reference for this paper
R. Fruchter, J. Gluck, "Adaptive Connections of Multitype Structural Descriptors in a TLS for Structural Analysis", in B.H.V. Topping, (Editor), "Artificial Intelligence Tools and Techniques for Civil and Structural Engineers", Civil-Comp Press, Edinburgh, UK, pp 141-148, 1989. doi:10.4203/ccp.11.6.1
It is the intent of this paper to present a model for adaptive connections of multitype structural descriptors in a training and learning system (TLS) for structural analysis. The development of this system uses the formalized methodology for learning from training experience throughout structural analysis. The system has a hierarchical architecture consisting of intelligent modules, suited for modeling structures. In the present version of the system, the knowledge that links abstraction levels is organized into two layers: a general heuristic knowledge layer and a specific-case knowledge layer. The latter is represented by structural descriptors and their corresponding training knowledge base (TKB). In this presentation the following issues will be emphasized: (1) knowledge representation of different types of structural descriptors, (2) acquisition of knowledge - in training rule (TR) form - related to the descriptor types, (3) learning of connections among structural descriptor TRs at a specific abstraction level - multitype connections - and the conditions in which they appear at different stages of structural analysis, (4) modeling a mechanism for adaptive connections among multitype structural descriptor TRs, and (5) effects of these connections on the structural response. The major goal of training and learning processes is to focus on critical modes of structural descriptors. Training and learning will assist the analysis process at structural module level and at training rule level. The central features of this system are represented by its capability to enhance its knowledge in a dynamic environment as well as its ability to determine, abstract, remember and use connections among multiple types of structural descriptor TRs at a specific abstraction level. The connections will adapt according to the changing demands of the structural analysis state. The development of this model is aimed to improve the performance of both structural analysis and TLS.

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £60 +P&P)