Computational Search in Architectural Design

The objective of this PhD research is to investigate how best to use computational search processes in the early phase of design. Search algorithms are implemented in combination with parametric models of different kinds of geometry, with the purpose of studying various building performances. The early phase of design and the nature of the design problem are stud-ied to get an idea of the kind of search process that would best accompany architects during this phase. Multi-disciplinarity and contrasting objectives are singled out as fundamental required characteristics of such a process. This leads to the proposal of multi-disciplinary studies into architectural shapes, both in the realm of complex curved geometry and in more traditional orthogonal forms. The search process itself is studied in its capacity to generate solutions in a multi-objective setting. The process of selecting and formulating search problems, the parametrization of geometric families for study as well as the selection criteria for outstanding solutions are all topics of discussion. A particular type of algorithm called Genetic Algorithms is implemented and studied in length. Search processes are proposed for 3 architectural design fields: structural, acoustic and energy design.