Evolutionary Computation and Artificial Life (202-1-5171), Semester A, 2007/8

Exercise 5

Implement the cellular programming algorithm with evolving architectures (C₁₂₉(1,d_i) and C₆₅(1,d_i)) to solve the synchronization problem, as shown in class.
Use biased (uniform) sampling to create initial (random) configurations, as explained in class.
Note: evolve connections more slowly than rules, e.g., rules evolve every C=300 configurations while connections evolve every C'=1500 configurations.

Submit:
1) Your program.
2) Plot of fitness versus time of three runs for two different-sized CAs: 65 and 129. Fitness = average fitness of all cells.
3) Space-time diagrams of some of your best evolved CAs' behavior.
4) Histograms of evolved connections.
5) Any improvements/tricks you've invented.
6) Any interesting conclusions you've reached.

Note: You can use and modify the ResultPad to create your diagrams.