Genetic Algorithm-Driven Optimization of Pattern for Parametric Facade Design Based on Support Position Data to Increase Visual Quality

Abstract

One of the most critical challenges for architects in façade design is providing an effective view from the indoors to the outdoors of a building for users, although the main role of the parametric façade is covering openings to control daylight and temperature. This study uses a genetic algorithm to optimize and evaluate the number and place of nodes and the position of supports required for a parametric façade based on the geometric patterns. Using the dataset with genetic algorithms is effective in reducing or replacing the nodes and supports of the façade. It also creates broader and irregular patterns just around the windows, which decreases the visual disturbance experienced by occupants. Accordingly, optimal building facade operation in terms of both building aesthetics and performance is important. The method used in this study, validated through three geometric grid patterns based on node positions, can be used to analyze dataset-incorporated patterns for potential irregular façade extensions. The nodes are considered by analyzing the cross-section optimization using the Galapagos program, and then data are obtained with Karamba based on reaction force, node force, and the deformation energy. The results show that among the three grid patterns, i.e., triangular, square, and hexagonal, the hexagonal grid is most efficient, exhibiting up to 60% lower reaction force, 40% lower node force, and 30% less deformation energy than the square grid pattern. The proposed GA also shows its effectiveness in enhancing the performance of parametric façades with patterns, thereby improving the occupants’ visual experience.