Illuminating Spaces: Deep Reinforcement Learning and Laser-Wall Partitioning for Architectural Layout Generation

TL;DR

This paper tackles the challenge of architectural space layout design by introducing laser-wall partitioning, a novel space-composition method that marries vector-based and pixel-based partitioning. It frames layout generation as a reinforcement learning problem within an $MDP$, optimizing with $PPO$ and evaluating via SpaceLayoutGym. Key contributions include one-shot and dynamic planning, on-light/off-light wall transformations, and identity-full versus identity-less wall assignments, demonstrated across six scenarios with strong geometric fidelity and high adjacency satisfaction. The approach yields diverse, functional layouts and offers architectural intuition, supported by an open-source simulator to enable further research and collaboration.

Abstract

Space layout design (SLD), occurring in the early stages of the design process, nonetheless influences both the functionality and aesthetics of the ultimate architectural outcome. The complexity of SLD necessitates innovative approaches to efficiently explore vast solution spaces. While image-based generative AI has emerged as a potential solution, they often rely on pixel-based space composition methods that lack intuitive representation of architectural processes. This paper leverages deep Reinforcement Learning (RL), as it offers a procedural approach that intuitively mimics the process of human designers. Effectively using RL for SLD requires an explorative space composing method to generate desirable design solutions. We introduce "laser-wall", a novel space partitioning method that conceptualizes walls as emitters of imaginary light beams to partition spaces. This approach bridges vector-based and pixel-based partitioning methods, offering both flexibility and exploratory power in generating diverse layouts. We present two planning strategies: one-shot planning, which generates entire layouts in a single pass, and dynamic planning, which allows for adaptive refinement by continuously transforming laser-walls. Additionally, we introduce on-light and off-light wall transformations for smooth and fast layout refinement, as well as identity-less and identity-full walls for versatile room assignment. We developed SpaceLayoutGym, an open-source OpenAI Gym compatible simulator for generating and evaluating space layouts. The RL agent processes the input design scenarios and generates solutions following a reward function that balances geometrical and topological requirements. Our results demonstrate that the RL-based laser-wall approach can generate diverse and functional space layouts that satisfy both geometric constraints and topological requirements and is architecturally intuitive.

Figures (6)

• Figure 1: Comparison of one-shot planning approaches using fixed infiltration rate with identity-less walls (top row) and decreasing infiltration rate with identity-full walls (bottom row)
• Figure 2: Illustration of off-light and on-light transformations in laser-wall partitioning
• Figure 3: Comparison of layouts generated using straight walls only (top row) and angled walls only (bottom row).
• Figure 4: Average rewards and episode lengths over training iterations .
• Figure 5: Comparison of space layouts generated using off-light (top two rows) and on-light (bottom two rows) transformations.