Dvosmerna integracija agentnih modelov z BIM za simulacijo energijske učinkovitosti stavb

This work addresses the significant “performance gap” in building energy simulation, which arises when fixed, standardized occupancy schedules are used in traditional BIM workflows. To improve prediction accuracy, this study presents a novel automated framework for bidirectional integration between Building Information Modeling (BIM) and Agent-Based Modeling (ABM). The framework uses a fivelayer workflow for closed-loop data exchange between a detailed Autodesk Revit model and the ABM platform. In practice, the BIM model is exported to ABM platform, where high-fidelity simulations generate dynamic, space-specific occupancy profiles from simulated occupant movements. The core contribution is a custom pyRevit add-in that automatically reintegrates these ABM-generated schedules into the Revit environment. By replacing uniform static schedules with dynamic ones, this process refines the internal load assumptions in energy simulations. Comparative analysis shows that incorporating the dynamic occupancy data has a measurable impact on simulation outcomes, notably changing metrics like annual Energy Use Intensity (EUI) and operational carbon emissions. The results confirm that automated, bidirectional BIM–ABM integration is both feasible and valuable for making building performance simulations more realistic. Ultimately, this framework offers a scalable approach for embedding human behavior dynamics into the design process, leading to more energy-efficient, occupant-centric buildings.