& Consulting 1,2,3{ekaterina.fuchkina|martin.bielik|sven.schneider}@ uni-weimar.de 4,5{t.ossenberg-engels| a.haemmerle}@oecc.de Graph-based representations of functional requirements (adjacencies, bubble diagram) are a common and useful method that supports architects in the conceptual phase of planning. However, the task of specifying the functional requirements through an adjacency graph can be challenging due to a quadratic growth of complexity in relation to the number of spaces. In turn, this increase of complexity challenges the designer searching for solutions that fulfill these functional requirements. There are systems that aim to address the difficulties related to graph-based space allocation. They, for instance, use fuzzy logic to weight the edges of a graph (i.e., specify relations between spaces) and spring systems (Newtonian gravitation model) to visually clarify the resulting proximity of all spaces according to the rules. Nevertheless, the problem of specifying large-scale adjacencies itself is omitted due to the assumption that such matrices are correctly filled in some previous steps.
| Authors | Ekaterina Fuchkina and Martin Bielik and Sven Schneider and Tobias Ossenberg-Engels and Alexander Hämmerle |
|---|---|
| Year | 2022 |
| Venue | Proceedings of the International Conference on Education and Research in Computer Aided Architectural Design in Europe |
| DOI | 10.52842/CONF.ECAADE.2022.2.039 |
| Source Database | Seed (CumInCAD) |
| Bridge-to-GNN | Category Seed |
| GNN Architecture | Not yet extracted from PDF |
| Graph Encoding | Not yet extracted from PDF |
| AEC Task | Not yet extracted from PDF |
| Cohort | Peak Adoption (2022) |
| Implementation Framework | Grasshopper / Rhino (heuristic — verify) |
|---|---|
| Primary Metric | Accuracy (heuristic — verify) |
| Quality Assessment | 1/6 · rigor: Low · Baseline: ✓ (heuristic — verify) |
| Frameworks / libraries | Grasshopper / Rhino |
|---|---|
| Quality (heuristic, 6-flag) | 0/6 · rigor: Low Code/Data: —Benchmark: —Baseline: —CV/Split: —Ablation: —Reproducible: — |
| Bibliometrically related to other seeds | This paper appeared in the Connected Papers neighbourhood of these corpus seeds: [55] 2021 · Collins · Assessing IFC Classes with Geometric Deep Learning on Dif… |
|---|
Part of the GML/GNN in AEC Systematic Review (PRISMA 2020) — 112 papers, 2020–2026