Knowledge-based parametric design: Applications on stairs

This study explores knowledge based parametric design, by rigorous development, documentation and evaluation of two applications in the domain of staircase design.; The first application is focused on the eight types of stairs presented in Palladio's First Book of Architecture, and involves a complete analysis of all types, as well as implementation in plan views, of the four types corresponding to oval and square stairs into a completed working system.; The second application in focused on "Narrow-U" egress stairs that meet the Uniform Building Code requirements, implemented completely for 3-plane symmetrical soffit designs with simple step geometry and depicted graphically in plan and section views.; Following a brief presentation of background information on knowledge based systems for design and stairs as a source of design knowledge, techniques for development and objective evaluation of knowledge bases for design are established. Based on this analysis, the proposed techniques for knowledge engineering are discussed and implemented for both applications. Finally conclusions are drawn on this process and on the impact of the environment, the expert system shell. The Top-Down system serves as the primary environment while an alternative expert system shell was developed for the AutoCAD environment to elaborate on specific technical issues.; A new generation of computing environments for knowledge based parametric design will partially automate some of the labor intensive procedures explored in this study and thus encourage and greatly facilitate large scale development of important applications with significant impact in the process of design.; Complete and working systems for knowledge based parametric design are valuable tools for the scholar and researcher of a particular domain of design, as well as the novice designer during formal training or experimentation. They provide support to the seasoned designer during exploration of a less familiar domain, while the confidence associated with the design output enables automated procedures for manufacture of the designed product with guaranteed results. But most significant, by enabling the development of the electronic treatise of the future, we are building the foundations of a system for archival and maintenance of design knowledge, thoroughly tested for internal consistency and lack of ambiguity.