Housing development planning increasingly demands quantitative methods that can balance its many potential environmental and economic impacts. Existing planning tools are hand-tailored to particular geographic regions, thus limiting adaptability, and use data models that don’t reflect the actual urban form.
Herein, we develop a generalizable mathematical programming framework for optimizing housing form and location at the spatial resolution of individual development sites. We show the adaptability of this framework by finding sustainable “gentle density” housing development plans in Toronto (Canada), Houston (USA), and Perth (Australia) under diverse sustainability objectives.
We first focus on minimizing embodied greenhouse gas (GHG) emissions from housing construction, and find that these three cities can theoretically house their projected populations with minimums of 3.0, 12.4, and 15.7 Mt CO2-eq of embodied GHG emissions (respectively) from housing architectural and structural materials. We examine the trade off between embodied carbon emissions and spatial accessibility (Toronto), flood risk (Houston), and climate damage risk (Perth), and find that these multi-objective optimization scenarios can increase the embodied carbon emissions up to 166% above the single-objective scenarios.
This work advances data-driven housing development optimization models by improving both spatial resolution and cross-city applicability.
