Search Search Close Search Close Search
Future Conscious Melbourne
Newsletter
Future Conscious Melbourne

Republish this Article

You may republish article text originally written for blueprint, unchanged, with credit to blueprint and a link to the original article. This permission does not cover photographs, illustrations, video or other third-party material. Check the stated licence or obtain permission from the rights holder before reusing those items.

The code below may include third-party material. Remove any items you have not independently confirmed you may reuse, and retain the required credits and licence links.

Parametric Design in Architecture

Published 23 August 2020 · Updated 13 November 2026

The gist: Parametric design links a building’s geometry to adjustable rules and relationships. Its value lies in testing alternatives and coordinating complexity, whether the result is curved or rectilinear. Software can expose useful choices, but the architect still defines the priorities, evaluates the trade-offs and takes responsibility for the design.

Parametric design describes a way of organising a design through relationships. Dimensions, angles, spacing or environmental inputs become variables in a model. Change one, and connected parts can update with it. The result may be a complex curved surface, a disciplined rectangular façade or a straightforward set of apartments.

This makes parametric design more useful as a method than as a period style. It has no single required material, silhouette or decorative language. A photograph can reveal repetition and variation, but it cannot, by itself, establish the process used to produce them.

Designing the relationships

McNeel’s Grasshopper provides a familiar example of the approach. Designers connect operations into a visual algorithm, defining how geometry responds when inputs change. Rather than drawing each element independently, they describe a system capable of producing and revising many related elements.

Consider a shading screen. Its panel spacing might respond to the structural grid, while panel depth varies with orientation. A revision to the grid can then flow through the screen model. The useful design work lies in deciding which relationships should exist, where exceptions are necessary and what the model is intended to test.

Variation is not the same as optimisation

A parametric model can generate alternatives; it does not automatically identify the best building. Optimisation adds a further step: alternatives are evaluated against chosen objectives and constraints. Those objectives may conflict. More daylight can bring more heat, a lighter structure can require more complex connections, and a highly varied façade can increase the number of components to procure.

For a project team, the question is whether the model makes those trade-offs visible. Useful outputs might include panel schedules, comparisons of shade or a coordinated geometric model. A convincing image alone says little about construction tolerances, maintenance access or the quality of the spaces inside.

Where computation meets a building

At Al Bahr Towers in Abu Dhabi, AHR developed a moving external screen inspired by the mashrabiya. Its units fold and unfold in response to the sun. The practice describes computational simulations and a shared geometry-and-performance manual used to carry the design into detailed development. Here, the relationship between a repeated component, its movement and environmental conditions is explicit.

Frank Gehry’s Guggenheim Museum Bilbao illustrates a related but different point. CATIA software helped translate a mathematically complex design into a structure that could be built. Digital coordination enabled the architecture; it does not follow that software independently invented the concept or selected an optimal form.

Read the model against the brief

A useful parametric process should make the project easier to understand and deliver. Ask which decisions remain adjustable, which assumptions have been fixed and how the proposed variations affect cost, comfort and construction. Keep the model tied to those decisions as the design develops.

The strongest application may be almost invisible in the finished building: a regularised family of façade panels, a more economical structural layout or shading that responds carefully to orientation. Complexity is justified when it resolves the brief, and simplicity can be an equally deliberate result.

The Great Court roof at the British Museum offers an earlier example of geometry organised around constraints. Working with Foster + Partners and Buro Happold, Chris Williams developed analytic and numerical methods to coordinate its form with architectural, structural and glazing requirements. The roof’s form emerged from coordinating those constraints through calculation.

Cover photograph: Great Court roof, British Museum, London, by DAVID ILIFF, via Wikimedia Commons. CC BY-SA 3.0. Resized; cropped in some previews.

Related reading: Modernist Architecture · Futurism and Futuristic Architecture.