Tuesday, September 15, 2020

[rswdvwpu] Skewed random cellular automata

Seed a deterministic pseudo random number generator with the states of cell and its neighbors.  Sample from the PRNG to get the cell's next state.  (Then, throw away the PRNG instance; it only gets used for evaluating that one cell.  Seed a new PRNG for the next cell and its neighbors.)

Consider sampling not from a uniform distribution but instead from some non-uniform distribution.  Some states being more common than other states is inspired by reality.  What are some good skewed distributions to sample from?  The binomial distribution is another non-uniform distribution.

Should a cell have a tendency to stay in the same state?  This would also reflect reality: inertia.

Consider defining a Markov chain among states.  What is a good way to define a nice Markov chain over 2^64 or more states?  Nice is an aesthetic quality, perhaps resembling or inspired by reality.  Things tending toward a sink state resembles some natural processes.

No comments :