Full Article:

Multi-objective problems usually employ conflicting objective functions, making the Simulation-Based Optimization process return a set of solutions. This study applies a clustering technique to analyze and characterize the solutions obtained in a school building optimization problem, maximizing daylight while minimizing energy consumption. We modeled the geometry using the Rhino + Grasshopper platform, following an existing building's characteristics. The parameters were the building's dimensions, openings' height, solar devices' and light shelves' reflectance, solar devices' distance from the facade, rotation angle, and depth of light shelves. We applied a clustering technique to group solutions according to their parametric similarities at the end of the optimization process. This approach made it possible to establish guidelines to support the designer's choice of the combination of parameters that best fits his purposes.