Process of evaluating and optimising architectural forms in multiple scenes using algorithmic designs

People’s impression of architecture is shaped by scenes from different angles and situations. However, despite their uniqueness, architectural forms created in recent years using algorithmic design methods, such as parametric design, tend to overlook the importance of appearance in different scenes. This study integrates different architectural appearances, along with architects’ intentions, from multiple viewpoints into the algorithmic design process. Specifically, we have established a multi-objective optimisation framework that treats the evaluation of architectural form’s appearance from each viewpoint as independent objective function. To assess how close the appearance of architectural form from each perspective is to what the designer intended, three items “silhouette”, “depth of form”, and “distribution of normal vectors on the constituent plane” were used. The significance and effectiveness of this process are examined through a case study that applies this method to a parametric architectural model with varying overall shape and surface pattern.