A better way to implement a "MiniMax" recursion

Question

I'm using the MinMax algorithm for a game, and since there are a lot possibilities the MinMax recursion takes way too long, even with "alpha-beta pruning"

My code looks some what like this:

min(state,depth,alpha,beta):
    if stopingCond: 
        return value

    for moves in allmoves:
        state.do(move)
        beta = min(beta, max(state,depth,alpha,beta) )
        if alpha >= beta: return beta
    return beta

max(state,depth,alpha,beta):
    if stopingCond: 
        return value

    for moves in allmoves:
        state.do(move)
        alpha = max(beta, min(state,depth,alpha,beta) )
        if alpha >= beta: return alpha
    return beta

I know that sometimes you can use for loops instead of recursion but I couldn't find a way to convert it.

If any one has a good idea I would be glad to hear it!

Thanks,

The game is a variation of poker, in each round there can be between 1 -to- 5 moves, and the max depth is 40 (5 O-Poker) — ifryed, Feb 25 '14 at 08:08
This question appears to be off-topic because it belongs on http://codereview.stackexchange.com — jonrsharpe, Feb 25 '14 at 08:47
In Python you have `def`. Also `moves` probably should be `move` — joaquin, Feb 25 '14 at 09:44
do you realize that depth 40 is huge with a branching factor of 2 it means 2**40 leaf so at reason of one evaluation by ms, it means 2.0**40/3600*1000/24/365 = 34e6 years — Xavier Combelle, Feb 25 '14 at 09:55

score 1 · Accepted Answer · answered Feb 25 '14 at 09:27

Generally the recursion in minmax can't be replaced by an equivalent iterative algorithm, and that's why there are optimization methods, like beta pruning, that try to improve performances by stopping some branches of the recursion tree as soon as possible.

It is possible that for your game, minmax (or, generally, a recursive algorithm) is not needed altogether and there could be other ways to find the optimal solution using other techniques, but without a precise set of rules is impossible to tell.

A better way to implement a "MiniMax" recursion

1 Answers1