Complex building performance analysis in early stages of design: a solution based on differential modeling, homology-based mapping, and generative design agents

Computational performance-based building design support faces a conflict. It is important to provide building performance feed back to the designer as early as possible in the design process. But many aspects of building performance are significantly affected by the design of the building’s technical systems, which are typically configured in detail only in the later stages of design. The challenge is thus finding a method to use detailed simulation tools even during the early stages of design when values for many of the variables for the building’s technical sub-systems are not yet available. In this paper, we offer a solution that involves differential modeling, homology-based mapping, and generative design agents. To demonstrate the feasibility of the proposed solution, we conclude the paper with illustrative examples of detailed performance analysis of complex buildings and their systems, in the early stages of design.