Arts & Technology

Artificial Intelligence: Evolutionary Programming

Last Update:

Cellular Automata

1. Cellular Automata & A-Life
   • Jon von Neumann's self-replicating cellular automaton (1955)
   • Conway's Game of Life (1967)
     • Cells on infinite checkerboard; each has 8 neighbors.
     • Cells are either alive or dead.
     • Alive cell
       • Will survive if 2 or 3 neighbors alive.
       • Will die if more than 3 (4,5,6,7,8), or fewer than 2 (1,0).
     • Dead cell
       • Remains dead unless exactly 3 neighbors alive: then born.
   • Demo applet ; Another demo applet [enlarge to full screen]
   • History of Life
     [Steven Levy, Artificial Life]
     • von Neumann's goal: self-reproducing automaton
     • Conway: simpler; and computer
     • Conjecture: the number of live cells does not grow without limit
     • Disproved by glider gun
       [A. K. Dewdney, The New Turing Omnibus]
2. Cellular Automata
3. Floys

Evolutionary Programming

1. Genetic Algorithms & Genetic Programming
   • Bit string (DNA) evolves
   • Program (individual) evolves
   • Mating strings
   • Mating programs
2. Dawkins' Biomorphs
3. Example: Blackjack
   • Rules of Blackjack/21: hit, stick, busted.
   • LISP: every program a tree.
   • [blackjack.lsp]
   • Performance? Approaches 50%, but highly variable.
4. Karl Sims Creatures
5. Genetic Programming for Art:
   • Music composition: GenBeBop
   • Art composition: Evol. Art page; Michael Gold

• Harold Cohen's Aaron