int alphabeta(int depth, int alpha, int beta, gameboard,player)
{
if(player == black)
curPlayer = blackplayer;
else curPlayer = whiteplayer;
if(depth == 0) return evalfunc();
generatePossibleMoves(possible_moves);
for (i = 0; i<possible_moves.size();i++)
{
if(timeleft() <= 0 || gameIsOver() != 0) break;
applyMove(possible_moves);
value = -alphabeta(depth-1,-beta,-alpha, gameboard, player);
myGame.undoMove(possible_moves);
if(value >= beta) return value;
if(value > alpha)
{
bestMoveTmp = i;
alpha = value;
}
if (globalDepth == depth) bestMove = bestMoveTmp;
}
return alpha;
}
Help with AB code
February 22, 2006 04:46 PM
Hello,
i have great problems with my source code to build an really
correct alpha beta algorithm.
Here is the alphabeta algorithm.
190
March 04, 2006 06:48 AM
Hi , looks like you are using negamax variation of minimax.Here is a small pseudo code for minimax with pruning
and here is pure minimax
and yes , switching players is necessary as based on the current players perspective , highest of all the children (MAX) or lowest of all the children (MIN) will be chosen.
Minimax(in game board, in int depth, in int max_depth, out score chosen_score, out score chosen_move ,int turn,int beta, int alpha)begin if (depth = max_depth) or (generate_moves() =0)then chosen_score = evaluation(board); return chosen_score; else moves_list = generate_moves(board); for (i = 1 to moves_list.length) do best_score = default; new_board = board; apply_move(new_board, moves_list); the_score=Minimax(new_board, depth+1, max_depth, chosen_score, chosen_move ,-turn,int beta, int alpha) if(alphas_turn) best_score = the_score; if(best_score > alpha) if( best_score < beta) assign best move ,best score & alpha; endif if(betas_turn) best_score = the_score; if(best_score < beta) if( best_score >alpha) assign best move , best score & beta ; endif endfor endifreturn chosen_scoreendand here is pure minimax
Minimax(in game board, in int depth, in int max_depth, out score chosen_score, out score chosen_move ,int turn)begin if (depth = max_depth) or (generate_moves() =0)then chosen_score = evaluation(board); return chosen_score; else moves_list = generate_moves(board); for (i = 1 to moves_list.length) do best_score =default; new_board = board; apply_move(new_board, moves_list); the_score = minimax(new_board, depth+1, max_depth, the_score, the_move,-turn) if (turn = max) if (the_score > best_score) best_score = the_score best_move=the_move endif endif if(the_score < best_score) best_score = the_score best_move=the_move endif endif endfor chosen_score = best_score; chosen_move = best_move;return chosen_scoreendand yes , switching players is necessary as based on the current players perspective , highest of all the children (MAX) or lowest of all the children (MIN) will be chosen.
"I think there is a world market for maybe five computers." -- Thomas Watson, Chairman of IBM, 1943
This topic is closed to new replies.
Advertisement
Popular Topics
Advertisement
