Abstract

It is widely acknowledge that by improving the building design process, its outcomes will improve as well. Therefore research efforts have been done to improve and to aid in the building design process. The advent of computers have revolutionize the Architecture, Engineering, and Construction (AEC) area; however, computers have been used mostly to increase productivity and to facilitate information exchange, but the building design process has remained essentially the same. The talk presents a research project that attempts to automate a simplified building design process by using existing and newly developed computational methods. The objective of the project is to aid the building design process by providing evaluated and optimised structural design variants of buildings in a systematic and automated manner, for which four approaches or strategies have been implemented. The implemented strategies were internally validated by performing sensitivity analyses to ensure that their procedures performed consistently; and they were externally validated to ensure that their results are transferable by processing generic buildings of 3, 8, and 18 levels. The implemented strategies are able to process entire buildings (instead of structural elements only), they prescribe optimal type, shape, and location of structural elements. Their main limitation is that the only consider structural engineering aspects; nevertheless they provide relevant structural engineering feedback at early stages of the building design process.