The layout of gymnasium directly affects environment performance. The current methods are insufficient to provide quantified decision support for gymnasium layouts in the early design stages (EDS). This study proposes a framework for optimizing the layout of gymnasiums using a multi-objective optimization (MOO) method based on genetic algorithms (GA) and neural networks. The study tested the framework using a community sports arena as an example, and the results indicate: the final optimized solutions achieved a maximum reduction of 11.1 % in cooling energy consumption (CE) and 3.3 % in solar radiation (SR) compared to the earlier generations, along with a 0.9 % improvement in thermal comfort percentage (TCP). This framework promotes the development of algorithm-driven methods for stadium layout design, while the prediction model based on RBF neural networks can simultaneously provide effective performance predictions for similar design outcomes.
Read full abstract