Generative Design Methodology and Framework Exploiting Designer-Algorithm Synergies

Designing is a problem-solving activity. The process is usually iterative: a solution is proposed, then analysed and tested until it satisfies all constraints and best fulfils the criteria. Usually, a designer proposes a solution based on intuition, experience, and knowledge. However, this does not work for problems they are facing for the first time. An alternative approach is generative design, where the designer focuses on iteratively defining a problem with its constraints and criteria in the form of a parametric computational model, and then leaves the search for the solution to the algorithms and their ability to rapidly generate and test several alternatives. The result of this approach is not only a set of solutions embedding implicitly the knowledge but also a model where problem-defining knowledge is quite explicit. The idea of the proposed approach is the exploitation of synergies between the designer and the algorithms. The designer focuses on problem definition and the algorithm focuses on finding a solution, showing that the capacity of the generative approach to replace the designer is limited. In the paper, we first present the framework of generative design, then apply the process to a case study of designing an efficient shading solution, and in the end, we present the results and compare them with the traditional approach. The approach is general and can be applied in other areas of engineering. It is relevant both to designers as well as software developers who are expected to take this approach further. More theoretical work is needed to study problem definitions as a form of knowledge representation in engineering.

Cette oeuvre a été publiée sous la license Creative Commons Attribution 4.0 (CC-BY 4.0). Il est autorisé de partager et adapter l'oeuvre tant que l'auteur est crédité et la license est indiquée (avec le lien ci-dessus). Vous devez aussi indiquer si des changements on été fait vis-à-vis de l'original.