I'm not familiar with it, but I'd assume most of them use the same mini-max, with alpha-beta prunning, or some similar algorithm, with small improvements, especially in the heuristic state evaluation function.

You'd learn more by studying a book on AI, and maybe trying to implement an engine yourself, iMO.

The difference in score between the top engines is not that big.