Intelligent Optimization of Spatial Parameters for Low-Carbon Campuses: A Study Based on Generative Design and Genetic Algorithms
In order to improve the computability and multi-objective coordination ability of the spatial form design of low-carbon campuses, in this study, GIS data processing, parametric generative design and NSGA-II are used. Eight variables such as green space ratio, canopy cover, permeable paving ratio and road network density were hybrid-encoded, while the net life-cycle carbon emissions, outdoor thermal comfort, and walkability were set as the optimization objectives by taking a typical open space at Chengdu East Campus as a study case. The results indicate that 36 non-dominated solutions were obtained from the algorithm. The comprehensive balanced solution resulted in a reduction of the net carbon emissions from 148.6 tCO₂e to 117.9 tCO₂e, an increase of the proportion of thermally comfortable area from 43.2% to 64.5% and an increase of walkability from 0.614 to 0.801. Spatial continuity, canopy supplementation and direct path connectivity allows for the synergic optimization of carbon reduction, environmental improvement and circulation efficiency within a limited site.