Knowledge-Based Assistance for the Analysis, Design and Optimization of Civil Structures

Abstract

An on-going project is described which has as its goal the development of an Al-based software package to assist in structural analysis and optimal design while explicitly treating uncertain risks. For preliminary design of a proposed structural system, it is desirable to search through a large design space to evaluate possible choices on the basis of multiple criteria, so that the most promising choice can be selected for detailed design. The goal of this project is to increase efficiency, fully integrate, and partially automate the design/build decision process by using AI- based methodologies such as knowledge-base expert systems, object-oriented programming, fuzzy reliability, multi-criteria decision theory and stochastic optimization. The methodology developed in this 11tudy allows the user to rapidly evaluate and improve a proposed conceptual design while taking into account the major factors affecting decisions related to design, construction, and operation in the presence of uncertain risk. These factors include not only structural engineering criteria, but also social, political, legal, and economic considerations. The first stage of this on-going project involved the development of a working prototype of an interfaced suite of knowledge-base expert systems consisting of analysis, design and optimization tools with user interactive features. The second stage of the project involves extending the capabilities of the software tools and introducing au explicit treatment of risk.