Generating the Optimal Pavilion Design through Sunlight-Driven Evolutionary Algorithms
Location: Poblenou, Barcelona

This project explores a generative design workflow applied to pavilion architecture in Poblenou, Barcelona, where sunlight is a critical environmental factor. Using evolutionary algorithms, we generate an optimized pavilion form that responds to solar exposure, ventilation, and material efficiency.

Inspired by the natural selection principle from John Holland’s genetic algorithms, this study applies Galapagos (Grasshopper) to simulate an automated, evolutionary design process. The goal is to explore how minimal human input can lead to high-performing architectural forms through digital tools.

Using Ladybug, we simulate the solar trajectory for Poblenou and feed it into a form-generation process driven by a combination of fitness criteria like sunlight exposure, ventilation, and geometry control.


To automatically generate and optimize a pavilion form using environmental data and a genetic algorithm to:

Explore a non-linear design process assisted by computation

Maximize sunlight acquisition and natural ventilation

Minimize material complexity and solar heat gain

Site & Climate Modeling

  • Establish site model in Poblenou using satellite or CAD data
  • Import solar radiation data via Ladybug

Parameterization in Grasshopper

  • Define genes:
    • Roof curvature (0.5–5 m)
    • Panel tilt angle (0°–45°)
    • Opening size (0.5–3 m)
    • Grid resolution (2–10 units)
    • Overhang depth (0–2 m)

Set Optimization Boundaries

  • Volume and footprint constraints based on site boundaries
  • Sun exposure goals: prioritize southern and eastern facades

Trigger Evolutionary Solver

  • Use Galapagos to evolve form towards best solar exposure
  • Fitness value = duration of sunlight exposure + airflow potential

Real-Time Monitoring

  • Adjust parameters live by monitoring evolution trends
  • Architect steps in only to fine-tune bounds, not to design directly

Select the Optimal Geometry

  • Choose the solution with highest fitness
  • Bake the final geometry and simulate sunlight for verification