Return
Room-level conditional graph neural networks for shear wall layout generation in high-rise buildings
H
Y
Y
DOI:10.1016/j.autcon.2026.107189.png)
Abstract
En 中文
• Room-level graphs encode spatial semantics and boundary feasibility. • FiLM conditioning enables discrete density-controlled wall generation. • Dual-stream training improves compliance with sparse paired labels.
Keywords:
Structural design automation
Shear wall layout
Graph neural networks
Conditional generation
Room-level representation
Journal
IF:
11.5
Papers:
6.1K
Citations:
4.2W
Organization
No organization information available
