Opossum

PAACADEMY offers advanced online courses on the Opossum Grasshopper plugin, teaching architects how to use machine learning-based surrogate modeling to optimize building performance efficiently.

Opossum is a Grasshopper plugin for machine learning-based optimization developed by Simon Watt, Mostapha Sadeghipour Roudsari, and others. Unlike evolutionary optimization tools such as Wallacei that use genetic algorithms, Opossum uses surrogate modeling, a machine learning approach where a predictive model of the design space is built from a relatively small number of evaluated samples, and optimized against rather than simulating every single iteration.

At PAACADEMY, Opossum is taught in Building Performance with Forma and Opossum by Olaf Olden, emphasizing its role in high-speed environmental analysis:

• Surrogate Modeling: Bypassing computationally expensive simulations by predicting performance based on machine learning data.
• Time Efficiency: Solving multi-objective optimization problems (like balancing daylight and thermal comfort) in a fraction of the time required by genetic algorithms.
• Sustainable Workflow: Integrating seamlessly with Ladybug and Honeybee to ensure climate-adaptive building massing and facade design.