Surrogate Models for Efficient Multi-Objective Optimization of Building Performance

Nowadays, the large set of available simulation tools brings numerous benefits to urban and architectural practices. However, simulations often take a considerable amount time yield significant results, particularly when performing many with models, as is typical in complex endeavors. Additionally, multiple objective optimizations metaheuristic algorithms have been widely used solve building optimization problems. most these processes exponentially increase computational correctly produce outputs require extensive knowledge interpret results. Thus, time-consuming rendered unfeasible requires specific methodology overcome barriers. This work integrates baseline multi-objective process used, validated energy tool. The goal minimize use cost construction residential complex. Afterward, machine learning techniques are create surrogate model capable accurately predicting Finally, different metaheuristics their tuned hyperparameters compared. Results show improvements results decrease up 22% total while having similar performance execution times 100 faster.