Building Information Modeling For The Built Environment Optimization

Abstract

This article presents measured performance results of a BIM-focused machine-learning (ML) optimization process for a 4000 sq ft mental-health rehabilitation center in central India. The process associated an Autodesk Revit BIM with Insight/EnergyPlus energy modeling, Radiance-based daylighting simulation, and a surrogate-assisted multi-objective genetic algorithm. In comparison with the baseline solution, the optimized plan resulted in significant improvements in daylight autonomy, thermal comfort, and natural ventilation and significantly cut down on annual energy consumption. Key outcomes are spatial Daylight Autonomy (sDA) +30 pp (50%→80%), Comfortable Thermal Hours +15 pp (75%→90%), Natural Ventilation Utilization +45 pp (40%→85%), and Energy Use Intensity −33% (150→100 kWh/m²·yr). The results show how tightly integrated BIM+AI workflows can expose non-obvious, climate-resilient design moves (orientation, glazing/shading mix, courtyard-driven stack effects) that collectively contribute to both sustainability and therapeutic quality in healthcare environments..