Welcome
Digital library of construction informatics
and information technology in civil engineering and construction
 

Works 

Paper w78-1993-2-331:
Managing construction knowledge in patterns: a neural network approach

Facilitated by the SciX project

Hegazy T, Moselhi O, Fazio P

Managing construction knowledge in patterns: a neural network approach

Abstract: Neural networks are AI-based computational tools with powerful capabilities of effective capturing and re-use of domain knowledge that are inherently implicit. This paper describes the modelling capabilities of neural networks with respect to construction problems, emphasizing the advantages associated with their representation of knowledge in the form of patterns. Several aspects related to proper management of knowledge are addressed for the purpose of developing practical and more reliable neural network models of complex construction problems. These aspects include: 1) problem structuring and patterns formation; 2) knowledge acquisition and data validation; 3) preparation and transformation of acquired data; and 4) analysis and interpretation of network state of knowledge. Guidelines pertaining to these aspects are provided along with considerations for modelling with noisy data and under high degree of uncertainty. The issues discussed are illustrated through a case study of a neural network for bidding decision support, developed based on knowledge acquired from contractors in Canada and the US. The case study demonstrates neural network modelling and illustrates the benefits gained through better management of acquired knowledge.

Keywords: neural networks; knowledge acquisition; construction; information technology; bidding strategy

DOI:

Full text: content.pdf (644,899 bytes) (available to registered users only)

Series: w78:1993 (browse)
Cluster:
Class:
Similar papers:
Sound: read aloud.

Permission to reproduce these papers has been graciously provided by the National University of Singapore. The assistance of the editors, particularly Prof. Martin Betts, is gratefully appreciated.

 

hosted by University of Ljubljana



includes

W78




© itc.scix.net 2003
Home page of this database login Powered by SciX Open Publishing Services 1.002 February 16, 2003