I recently saw a video about Wolframs "new kind of science" and started to read the book, as a hobbyist in regards to cellular automata, that always wondered if cellular automata could be translated into a more basic graph structure I am very interested in wolframs idea about updating graphs.
Now I am also quite a fan of the idea of selfevolving code and I always wondered if one could not write a genetic algorithm for code, but alas I never had the expertise to programm an experiment for this, as I quickly realized that such an algorithm would need to be taught how to navigate the graph of a code's syntax.
Now if we had a ruleset for updating graphs that was computational universal, everything could be represented in the form of a graph and it should be much easier and natural to write a genetic algoritm able to operate on itself. If such a ruleset was causal invariant, simulation would be even easier.
Such my question: what is simplest causal invariant and computational universal ruleset( if there are rulesets able to model the universe they also should be computational universal right?)