Space-level circulation design rationale refers to the underlying logic and principles that guide the arrangement of pathways, connections, and spaces within a building to facilitate movement and accessibility. Capturing and reusing this rationale is crucial for designers, as it enables the efficient adaptation of existing designs to new contexts, preserving both functionality and intent while reducing the time and effort required for rework. Traditionally, reasoning about circulation design has relied on manual interpretations, conceptual sketches, and designer intuition, often resulting in incomplete or ambiguous representations of spatial relationships. These approaches, while creative, lack the systematic rigour necessary for ensuring consistency and adaptability across diverse design scenarios. This thesis explores the application of graph-based analysis techniques, including graph clustering algorithms and centrality-based node analyses, to reason space-level circulation design rationale. The method computes space-level topological relationships by leveraging IFC schema relationships and utilizing a voxel-grid-based approach, establishing graph-based representations of IFC models. These representations provide a systematic means of understanding space-level organization and connectivity. This research demonstrates the potential of integrating graph-based concepts into building design and verification processes, highlighting the effectiveness of graphs for the representation and analysis of circulation entities.     «

Space-level circulation design rationale refers to the underlying logic and principles that guide the arrangement of pathways, connections, and spaces within a building to facilitate movement and accessibility. Capturing and reusing this rationale is crucial for designers, as it enables the efficient adaptation of existing designs to new contexts, preserving both functionality and intent while reducing the time and effort required for rework. Traditionally, reasoning about circulation design has...     »