20061002
WEEK FIVE
When search spaces (or "adaptive landscapes") were first postulated in biology in the 1930s, they were thought to be prestructured by a single equilibrium, a kind of mountain with one peak, which selection pressures forced the probe head to climb. According to this schema, the top of the mountain represented the point of maximum fitness, and once a population had been driven there, selection pressures would keep it locked into this optimal equilibrium. However, recent explorations of adaptive landscapes, using sophisticated computer simulations, have revealed that these search spaces are anything but simple, that they may comprise many mountains of different heights (local optma), clustered in a variety of ways, the valleys and peaks related not directly to fitness but underlying dynamical stable states. - Manuel De Landa
Probe
Probe the space of Math variation. Dive into the studio pool of 33,000 surfaces (generated from 11 default surfaces) and grab three that reveal similarities to the spatial moments that you have identified and built previously. Start re-running the respective parametric changes in smaller increments to grow three clouds of finer variation. (math models)
Digitize
Digitize your physical models. Numerically define points in x-y-z space; high points, turning points, control points, end points. Unwrap, unfold your foamy and paper models to measure and record dimensions. In Rhino build Nurb surface models derived from your foamy models and Polygonal models from your paper models. Operate economically with control points and polygons. (manual models)
Exchange
Nurture an intensive exchange between math models, manual models and physical models. Example: Derive surface curves from a math model, manipulate the curves and generate a next generation manual model. Unfold the manual model, print the wire frame planar and prepare sheet material accordingly, then build a next generation physical model. Fine tune parametric variation of math model to approximate physical model. Record your steps.
Grow
Based on the continuous interlocking field study (foam, paper, pixel pattern) develop a combinatory logic and grow tectonic fields for the rhino and physical models. Probe and learn from different arraying techniques (along line, along surface). Focus on your cells' continuous variation. Parametric variation across the field generates change beyond a mere scalar shifts. Build the field testing boundary conditions and extreme values. Draw the field. Write the script.
Exhibit
Develop a concept for displaying the artifacts. The tectonic field can stage, frame or integrate the items; it could merge into the objects at certain moments, wrap around them or remain at a safe distance to them. You could eliminate the artifacts altogether - an extreme position - if the structure itself best celebrates or informs about the artifacts. Every approach to exhibiting the artifacts will invite a set of demands into your structure. Draw a catalog of responses.
Read
Manuel De Landa : Species and Ecosystems
.: Jonas 4:00 PM
|