Responsive Architecture through Parametric Design

How Parametric Systems Introduce Diversity, Adaptability, and Responsiveness into the Architectural Design Process

Ferda Kolatan

University of Pennsylvania

su11 architecture+design

In recent years generative systems in nature have become a source of inspiration for a number of architects. Rather than viewing design as a top-down, linear, and goal-oriented procedure, they seek to understand underlying rules and principles of natural systems, which produce structural order and material organization of high complexity, efficiency, and beauty. Our own research investigates how parametrically controlled systems can be applied to architectural design in order to achieve more responsive and adaptive modeling techniques. While working with the software GC (generative components) we also found that a rule-based approach to design introduces a whole new set of opportunities, which as designers we have the responsibility to further explore and develop.

Parametric software in architecture is often viewed as merely an output tool, which allows for complex shapes to be translated into fabrication and production modes. As such, this type of software marks the final steps of the design process while often more “intuitive” (graphic interface) modeling software is used for the initial design of the project. However, if used as an input tool, parametric setups promise an entirely novel approach to the field of architecture, which can be identified through a set of specific characteristics such as the following:

A parametric design system is based on a hierarchy of variable-controlled dependencies. Each active variable causes the overall system to change its behavior and thereby generate variations without losing the overall coherence and integrity of the model. This is an incredibly powerful tool, which allows designers to test a range of possible applications within a controlled design environment.

Which variations are “fitter” to solve a particular design problem? Parametric models can be fine-tuned toward specific environmental pressures. Versions can be tested toward particular performance values. Different regions within the same model can be programmed and constrained individually in order to respond more sensitively. Environmental changes that occur over time at the same location can be addressed by layering functional zones and controlling them through targeted variable manipulations.

A parametric model maintains its ability to change (openness) throughout the entire design process. Individual components can be linked in ways that allow constant and immediate feedback throughout the model regardless of scale and hierarchy. The design of the network of dependencies becomes a crucial part of the overall design by highlighting certain relationships over others and thereby allowing for specific local changes to impact the model’s global scale and behavior.

Both in our practice as well as with my students we have worked on a number of design projects, which have begun to explore the above issues. Wolfram’s work, specifically his book A New Kind of Science, has been of great significance for our work and an inspiration to our thinking.

[presentation materials]

Created by Mathematica  (May 11, 2006)