Algorithmic Form Finding for Solar Design

How do you design a building that maximizes square footage while preserving sunlight access for your neighbor across the street?

That was the question motivating this study for an institutional client. The two driving factors — maximum developable area and sunlight preservation for a greenhouse and patio across the street — conflicted with one another. Improving one made the other worse, so a trade study was needed.

Our approach involved sunlight simulation, site envelope generation, a genetic algorithm, and a data-driven feedback loop that fed into the optimization problem. But the design criteria also needed judgement imposed: how do you translate between the value of an additional square foot of developable space for the building and an additional hour of sunlight access for the greenhouse? If you want the algorithm to converge on a single answer, you have to decide up front.

But this is where it helps to understand that trade-off problems like this, multi-objective optimizations, have a range of “right” or Pareto efficient answers. These show up along a Pareto front or curve, where no combinatorial solution can be outperformed by another. If the quantities you’re trading are more discrete, there's a more manageable number of solutions; as they get more continuous, the number of solutions greatly increases. For example, if each additional square foot of building area has value, then there will be a large number of very similar solutions, but if that additional floor area is only useful if you can add another classroom, then there will be fewer, more distinct solutions. So the Pareto curve is less of a smooth curve and instead has steps.

This animation keeps a lot of the details under the hood and is the last of many runs, which is designed to converge on a single solution. Being iterative makes it easier to inject human judgment and guidance into the feedback loop and solution space. The final answer here is not intended to move directly into construction, but to inform the design team and client about the trade space, helping everyone better understand the design problem and more quickly land on a outcome that responds to the client's goals.