I haven't posted so much this period 'cause I'm going to graduate on 18th of July. Thesis title is:
Computational Morphogenesis in Architecture - Free-Form Spatial Structures Optimization.
I worked in particular on spatial frames (like the great court of the British Museum - Foster + Buro Happold), simulating their behaviour in Rhino with a mesh. In a multi objective aim that concerns geometry and statics I developed algorithms to drive lengths between mesh vertices to a given database of measures, and then to perform mesh state of stress under gravity loads (for this second step I have linked Rhino and Ansys by a memetic algorithm in RVB).
The smaller the database, the roughter the result structures look. The difference from the 'original automatic made mesh', in the case of a hyperbolic paraboloid, is shown below.
Here are 3 benchmarks that I used to test different algorithms. Pink lines represent lengths driven to database measures.