U-Net-Accelerated Quality-Diversity Optimization for Climate-Adaptive Urban Layouts

2026-06-03

AI Summary

This study addresses the challenge of optimizing climate-resilient urban layouts by balancing building density and cool-air ventilation, a task traditionally hindered by the high computational cost of high-fidelity physical simulations that limits design-space exploration. To overcome this, the authors propose embedding a U-Net deep learning surrogate model—leveraging its spatial inductive bias—into an offline MAP-Elites quality-diversity optimization framework. Trained solely on a one-time Sobol random sampling dataset, the surrogate achieves highly accurate prediction of microclimate responses (R² = 0.996) without requiring costly active sampling. The approach maintains strong rank-order consistency in layout fitness (Spearman’s ρ = 0.994) and enables the rapid generation of thousands of diverse, high-performing urban design alternatives within minutes, substantially enhancing both optimization efficiency and generalization capability.

Abstract

Optimizing urban layouts for climate adaptation requires balancing building density with cold-air ventilation. Because physics-based climate simulations are computationally expensive, planners typically evaluate fewer than ten manual designs. \gls{qd} algorithms offer a way to systematically illuminate the design space, but they require surrogate models to be practical. In this paper, we replace a slow, regulatory physics simulator with a spatial deep-learning surrogate (U-Net) inside an offline MAP-Elites loop. We systematically compare this spatial approach with a traditional \gls{gp} surrogate across different training-data strategies (quasi-random Sobol sampling vs.\ active \gls{qd} bootstrapping). Our results reveal that scalar \gls{gp} surrogates fail catastrophically when trained on random samples, requiring expensive, actively generated \gls{qd} archives to generalize. In contrast, the spatial inductive bias of the U-Net allows it to learn the underlying physics mapping robustly ($R^2 = 0.996$), completely independent of the training data source. This allows offline \gls{qd} optimization to achieve highly accurate fitness rankings ($ρ= 0.994$) using only a one-time batch of random training samples. The resulting pipeline, deployed in the open-source OpenSKIZZE tool, generates thousands of diverse, climate-evaluated building layouts in under ten minutes.

Problem

• climate-adaptive urban layouts
• quality-diversity optimization
• computational efficiency
• surrogate modeling
• urban design

Innovation

• U-Net
• Quality-Diversity Optimization
• Surrogate Modeling
• Climate-Adaptive Urban Design
• MAP-Elites