INTRODUCTION
We approached our parametric project with a top-down methodology, emphasizing the importance of urban, climatic, social, and economic parameters. These factors, alongside the needs for project development, are crucial when discussing architecture that focuses on the reuse and recycling of existing structures in both established and developing contexts.
CONCEPT & STAGES
The algorithm is initiated on a designated plot with a defined access road, but the flexible script can adapt to any location.
Starting with the premise of working on existing buildings, typically characterized by regular reinforced concrete structures common in outdated residential or commercial buildings in Europe, we implemented horizontal and vertical expansions. This approach introduces service areas around the existing structure.
Sunlight as an environmental parameter is used to optimize heights and vary facades, incorporating terraces and green roofs. The old structures house administrative functions, while the new additions are dedicated to residential purposes, the towers, and various services within the plot.
PSEUDOCODE
The workflow is divided into preprocessing steps, building generation, and post-process data outputs for evaluation.
- Preprocessing: Extracts information about the existing structure, such as grid size and column and beam dimensions, and refines topography and massing for the new structure.
- Generation: Relies on the geometrical shapes of existing and new structures and spatial demands like FAR and room height.
- Postprocessing: Involves structural analysis to optimize material usage and reduce displacements and stiffness, and environmental analysis to optimize sunlight access.
Common Challenges:
- Adapting the overall generation to plot size and topography.
- Placing functions within the building according to existing and new structures.
Group Challenge 1:
- Conducting more than a quantitative assessment.
- Optimizing the building towards qualitative goals such as social, sports, leisure values, nature, mobility.
Group Challenge 2:
- Adapting facade design based on context.
- Automatically adjusting window sizes based on solar radiation analysis, using a distribution of four different facade modules.
FLOORPLAN GENERATION
Topological map
The spatial arrangement is governed by a topological map, which defines the connections between modules based on adjacency requirements and functional relationships. This ensures that each configuration aligns with programmatic needs while maintaining logical and efficient spatial flows for every generations. For instance, circulation modules act as connectors, strategically linking residential and office spaces to shared amenities and services.
Wave Function Collapse for Floorplan Generation
At the heart of this project lies a sophisticated implementation of the Wave Function Collapse (WFC) algorithm, adapted to generate contextually responsive floorplans. The process begins with a catalog of modules, categorized by function—residential, office, social, services, circulation, and more. These modules serve as the building blocks for the design, allowing for high adaptability and a modular approach.
The modular system is further enhanced by its parametric flexibility. Modules are designed to adapt to changes in grid size, with all proportions and dimensions dynamically parameterized. This adaptability allows the design to respond seamlessly to site-specific constraints, such as varying plot sizes or urban densities, ensuring that the generated solutions remain scalable and context-sensitive.
This WFC-based approach is a key component of the broader generative design workflow. The script iterates through multiple design options, automatically analyzing metrics such as structural efficiency, carbon footprint, and spatial performance. With live feedback, each iteration refines the design to achieve an optimal balance of sustainability, functionality, and aesthetic quality.
Data
Performance data, including environmental metrics, spatial efficiency, and assigned qualitative performance attributes, are continuously updated in real-time, providing immediate feedback. These attributes align with our group challenge to integrate both measurable outputs and design quality. This iterative approach ensures that each design balances contextual responsiveness with sustainability and functional performance. Key metrics include:
- FAR (Floor Area Ratio): Detailed calculations with variations.
- Program Ratio & Percentage: Clear allocation of spaces.
- Qualitative Assessments: Addressing usability and design quality.
- Structure: Evaluation with Karamba
- Environmental Data: Solar analysis and facade types
- timing of the script
VISUALISATION
Singapore/ FAR 6
Rendered with ComfyUI, the design leverages Flux.1 Dev with LoRA Depth and Canny to follow the design’s shape. Materials like wood and glass provide warmth and transparency, while green terraces create a serene communal space. The facade merges old and new building elements seamlessly.
Reykjavik / FAR 2
With the same technic of vizualisation, the tailored prompts transport the concept to Iceland. The design features high-rise buildings that harmonize with Icelandic culture and urban landscape, using sleek modern materials to meet Reykjavik’s winter aesthetics.
By combining advanced computational techniques with architectural expertise, this project exemplifies how generative design can push the boundaries of traditional workflows. It not only accelerates the design process but also ensures that the outputs are environmentally and contextually aligned.
The integration of algorithms like Wave Function Collapse demonstrates the potential of computational design to create adaptive, high-performance architecture—bridging the gap between data-driven processes and human-centric design.
References
https://monoceros.sub.digital/monoceros-user-manual/index.html
