A Case Study on Evaluating Genetic Algorithms for Early Building Design Optimization: Comparison with Random and Grid Searches

In early-stage architectural design, optimization algorithms are essential for efficiently exploring large and complex design spaces under tight computational constraints. While prior research has benchmarked various optimization methods, their findings often lack generalizability to real-world, domain-specific problems, particularly in early building design optimization for energy performance. This study evaluates the effectiveness of Genetic Algorithms (GAs) for early design optimization, focusing on their ability to find near-optimal solutions within limited timeframes. Using a constrained case study, we compare a simple GA to two baseline methods, Random Search (RS) and Grid Search (GS), with each algorithm tested 10 times to enhance the reliability of the conclusions. Our findings show that while RS may miss optimal solutions due to its stochastic nature, it was unexpectedly effective under tight computational limits. Despite being more systematic, GS was outperformed by RS, likely due to the irregular design search space. This suggests that, under strict computational constraints, lightweight methods like RS can sometimes outperform more complex approaches like GA. As this study is limited to a single case under specific constraints, future research should investigate a broader range of design scenarios and computational settings to validate and generalize the findings. Additionally, the potential of Random Search or hybrid optimization methods should be further investigated, particularly in contexts with strict computational limitations.

Convolutional versus Dense Neural Networks: Comparing the Two Neural Networks Performance in Predicting Building Operational Energy Use Based on the Building Shape

A building self-shading shape impacts substantially on the amount of direct sunlight received by the building and contributes significantly to building operational energy use, in addition to other major contributing variables, such as materials and window-to-wall ratios. Deep Learning has the potential to assist designers and engineers by efficiently predicting building energy performance. This paper assesses the applicability of two different neural networks structures, Dense Neural Network (DNN) and Convolutional Neural Network (CNN), for predicting building operational energy use with respect to building shape. The comparison between the two neural networks shows that the DNN model surpasses the CNN model in performance, simplicity, and computation time. However, image-based CNN has the benefit of utilizing architectural graphics that facilitates design communication.