How to Evolve a Cellular Automaton   no comments

Posted at 12:47 pm in Uncategorized

This week, I’ve been pushing on with “Complexity: A Guided Tour”, and continued attending the Complexity lectures.  Their subject matter are converging towards “in what sense do real-world distributed complex systems compute?” (e.g. ant colonies or the Web).  This is very relevant to my key theme of collaborative problem solving. It builds on last week’s introduction to cellular automata.  Most fascinating was an experiment that used genetic algorithms to evolve cellular automata to perform global analysis despite being highly distributed.  The results was the emergence of Feynman-diagram like particles, operating at the abstracted level equivalent to the programming level of a traditional von-Neumann-style computer.  Obviously, computation has been a key theme throughout, so there’s a happy overlap with the theory of computation COMP6046 lectures.  I’ve enjoyed the introduction to information theory and Shannon entropy, and have chosen Comp Thinking coursework on encryption in order to build on that.  The other key ingredient this week has been thermodynamics, and this has been a happy stroll back to my time in Physics.   Above all the Darwin-like obvious-yet-brilliant central observation of Boltzmann that things will always tend toward the more common state – hence the mighty 2nd law of TD.  So, several really engaging themes have been weaving together.  Looking forward to what comes next.

Written by Jack on November 23rd, 2010

Leave a Reply