Content area
With the continuous growth of urbanization and the ambitious decarbonation objectives to tackle climate change, new solutions are needed to optimize energy management in dense urban landscapes. This research advances MILP-based (Mixed Integer Linear Programming) optimization for large-scale urban applications by integrating a Warmstart technique that reduces computational time and increases the number of buildings considered. The problem is restructured to (1) solve the optimization problem with a more precise solution, (2) distribute the renovations across the various periods of the optimization problem, and (3) initialize the problem with the created solution. By significantly improving computational efficiency, the OptoBAT method presented in this paper can now account for a larger and more representative sample of urban building clusters, moving beyond simplified medoid representations to incorporate more granular spatial data. To evaluate the impact of these enhancements, this study compares the results of energy optimization for a French metropolitan case study before and after integrating the updated MILP methodology. The findings reveal improved model scalability, a reduction in computational demands by more than 75%, and potentially more accurate energy optimization outcomes. This research contributes to the field by advancing MILP-based optimization for large-scale urban applications.