I am a new learner of AI and I understand expectimax way of working, however, I got a new term called expectiminimax. So, does expeciminimax differ from expectimax? Or they are the same terminology?
If no, then what are the main differences between them?
I'm still learning too, but I'll share what I've learned so far.
In simple terms they are the same, with the only difference being expectimax is for single-player stochastic games, while expectiminimax is for two-player stochastic games.
You see, expectiminimax is a variation of minimax. In minimax, there are two players, MAX who wants to maximize utility for itself and MIN who wants to minimize the utility for MAX. Now in stochastic games, the utility is also affected by probabilities. So in expectiminimax, MAX wants to maximize its expected utility while MIN wants to minimize the expected utility for MAX. One example of game where this can be used is Backgammon where there are two competing players and both roll dice.
Meanwhile, in expectimax, there is only one player, namely MAX. Since the outcome of its action are nondeterministic, MAX would want to maximize its expected utility. One example of game where expectimax is used is 2048.
