Some extensions to Cellular Automata include adding intelligence or learning capabilities to the automata, in which case they can simulate intelligent agent behaviour.

Swarm intelligence,

in some ways, may be viewed as an extension of cellular automata, where they deposit evanescent pheremone trails on the return journey after a successful trip. In effect, they are providing a time-decaying positive feedback mechanism to reinforce the better outcomes for randomised searches for local optima and by extension will tend towards a global optimum - through an attempted optimal combination of local optima. The time-decaying feedback allows the system to adapt dynamically to changing circumstances and avoids getting stuck in local minima.


Another option is to view them as quasi-optmised probabilistic searches over space and time, which avoid getting stuck in local optima.