Multi-objective optimization design of residential area based on microenvironment simulation

In current research on old communities, less attention is given to the comfort and greenhouse gas emissions of the indoor and outdoor physical environment. Only a few studies focus on optimization strategies for these environments in old communities. Thus, the absence of suitable theoretical guidance in specific engineering practices can lead to issues. Renovated old communities still face challenges such as overheating in summer, overcooling in winter, and high energy consumption. In this article, based on the optimization of low- and medium-rise old communities in five typical cities in China, corresponding strategies such as geometric and envelope design parameters are proposed in terms of building greenhouse gas emission and indoor and outdoor thermal environment comfort. Several methods are employed in this study, including parametric simulation using Grasshopper with the Ladybug and Honeybee plugins for data collection. This research conduct sensitivity analysis using Random Forest Regressor (RFR), Gradient Boosting Regressor (GBR) and Decision Tree Regressor (DTR) to explore the relationship between design variables and environmental objectives. Predictive modeling with ensemble models and multi-objective optimization using the NSGA-2 algorithm help identify optimal design parameters for varied urban climates. The results demonstrated indoor and outdoor comfort of communities can be improved to a greater extent while reducing carbon emissions. Quantitatively, compared to the baseline model, all cities except Wuhan can improve the year-round comfort time of the outdoor thermal environment by more than 50%. Besides, communities in hot summer and warm winter regions (e.g. Shenzhen and Nanning) can reduce their annual carbon emissions by more than 300t, and communities in hot summer and cold winter regions (e.g. Hangzhou, Chengdu and Wuhan) can reduce their annual carbon emissions by more than 100t. The research results will serve as a scientific basis for energy saving and emission reduction and provide decision-making methods for the development of old communities. •Sustainable microenvironment evaluation system established.•RFR, GBR, and DTR compared for sensitivity.•Latin Hypercube Sampling enhances meta-models.•Optimal city environmental performance compared with baseline models.