Boxes (Round 2)

Then we had to make fifty boxes! I used a survey and a genetic algorithm over the course of four generations to determine which boxes came out on top.

Boxes (Round 2)

More Boxes (Early Generation)

Here's an earlier generation of the boxes project for interface class. Most of these were "killed" later by the genetic algorithm I used to generate more boxes.

More Boxes (Early Generation)

More Boxes (Later Generation)

This is a later generation of boxes. The losers are on the right. My genetic algorithm evaluated each box based on preferences I got from people in surveys. The genome of each box contained its color, size, opacity, and saturation. Clearly, the transparent ones were not the fittest. The red ones were quite successful, though.

More Boxes (Later Generation)

More Boxes (The Dustbin of History)

Once I was finished with the boxes, the ones that never made the cut were crushed. Think of my foot as the Permian-Triassic extinction event.

More Boxes (The Dustbin of History)

The Algorithm

This is the python script I used to determine the winners and losers among my box designs. Each box had a genome that encoded its color, saturation, size, and opacity. Boxes with more votes were allowed to mate. The rest were crushed. Circle of life.

AttachmentSize
box_gene.py.txt2.12 KB