Innovative Hospital Floor Planning Using Graphormers

Designing a floor plan requires extensive planning and expertise. Particularly when hospital floor plans are being designed, it is crucial to optimize them so that navigating between specific rooms is minimal. Our aim is to generate a model that automates floor plan generation for hospital layouts. In our approach, we use Multi-Model AI Architecture for automating hospital layout design. Initially, a CNN model is trained to predict the node center in a given input space that represents rooms. Each node represents its embedding-information about the room type, relationships, functional proximity, etc. A graphormer model examines these nodes and captures both local and global attention mechanisms to predict the node connections and relations. Finally, a GAN is used to generate a visual representation of the finalized layout. This proposed approach could provide more efficient generation of hospital floor plans while considering factors crucial for increasing the overall workflow. Thus resulting in a functional layout that improves both patient and staff experiences, which leads to efficient hospital operations.