Abstract:The architectural environment is expanding into digital, virtual, and informational dimensions, introducing challenges in virtual-real space interaction. Traditional design methods struggle with real-time interaction, integration with existing workflows, and rapid space modification. To address these issues, we present a generative design method that enables symbiotic interaction between virtual and real spaces using Mixed Reality (MR) and Generative Artificial Intelligence (AI) technologies. We developed two approaches: one using the Rhino modeling platform and the other based on the Unity3D game engine, tailored to different application needs. User experience testing in exhibition, leisure, and residential spaces evaluated our method's effectiveness. Results showed significant improvements in design flexibility, interactive efficiency, and user satisfaction. In the exhibition scenario, the Unity3D-based method excelled in rapid design modifications and immersive experiences. Questionnaire data indicated that MR offers good visual comfort and higher immersion than VR, effectively supporting architects in interface and scale design. Clustering analysis of participants' position and gaze data revealed diverse behavioral patterns in the virtual-physical exhibition space, providing insights for optimizing spatial layouts and interaction methods. Our findings suggest that the generative AI-driven MR method simplifies traditional design processes by enabling real-time modification and interaction with spatial interfaces through simple verbal and motion interactions. This approach streamlines workflows by reducing steps like measuring, modeling, and rendering, while enhancing user engagement and creativity. Overall, this method offers new possibilities for experiential exhibition and architectural design, contributing to future environments where virtual and real spaces coexist seamlessly.

KOALA - Developing a Generative House Design System with Agent-Based Modelling of Social Spatial Processes

Integration of Agent-Based Modelling of Social-Spatial Processes in Architectural Parametric Design

A emotion development agent model based on OCC model and operant conditioning

An agent-based model of public space use

Building Occupant Transient Agent-Based Model – Movement Module

Enhancing Interaction in Virtual-Real Architectural Environments: A Comparative Analysis of Generative AI-driven Reality Approaches

Simulation and Visualization of Energy-Related Occupant Behavior in Office Buildings

Influence of interior layouts on occupant energy-saving behaviour in buildings: An integrated approach using Agent-Based Modelling, System Dynamics and Building Information Modelling

Modelling Office Energy Consumption: An Agent Based Approach

Sculpting the Space: A Circulation Based Approach to Generative Design in A Multi-Agent System

Multi-agent based energy model for domestic duildings based on occupant behavior

The Adaptive Architectural Layout: How the Control of a Semi-Autonomous Mobile Robotic Partition was Shared to Mediate the Environmental Demands and Resources of an Open-Plan Office

A novel approach for occupants' horizontal and vertical movement modeling in non-residential buildings using Immersive Virtual Environment (IVE)

Agent-Augmented Co-Space: Toward Merging of Real World and Cyberspace

Agent-based post-earthquake evacuation simulation to enhance early-stage architectural layout and non-structural design

Exploring Communication in Collaborative Design: Co-Operative Architectural Modelling

Embracing AWKWARD! Real-time Adjustment of Reactive Plans Using Social Norms

Towards occupant-centric simulation-aided building design: a case study

Enhancing Spatial Configurations in Informal Learning Spaces: an Early-Phase Design Decision-Making Framework

A Generative Design Method for the Functional Layout of Town Planning Based on Multi-Agent System