本文详细介绍了Alpha-Beta剪枝搜索算法的工作原理及其实现过程，包括基本的Alpha-Beta搜索流程、如何通过剪枝提高搜索效率，以及如何在实际游戏中应用此算法来寻找最优解。
// basic Alpha-Beta search;
move bestmove;
int alphabeta(int depth, int alpha, int beta, bool isMyTurn) {
if (gameOver || depth == 0) {
return evaluation();
} generateLegalMoves(); // you may change the search order to make the search faster for (each move m) { do move m; score = -alphabeta(depth - 1, -beta, -alpha, ! isMyTurn ); if (score > alpha) { alpha = score; if(at the Level 1) bestmove = m; } undo move-m; if (alpha >= beta) { break; } } return alpha;
}Initial call : alphabeta(depth, - infinite, infinite, isMyTurn);
不理解上面的代码的话，看看下面的极小极大搜索
极小极大搜索
//极大搜索int Max(int depth){ int best = -INFINITY; if (depth == 0) { return evaluation(); } generateLegalMoves(); for (each move m) { { do move m; val = Min(depth - 1); // call the Min search undo move m; if (val > best) { best = val; } } return best;} int Min(int depth) // Min search{ int best = INFINITY; if (depth <= 0) { return evaluation(); } GenerateLegalMoves(); for (each move m) { do move m; val = Max(depth - 1); // call the Max search undo move m; if (val < best) { best = val; } } return best;}
负极大搜索
极小极大搜索可以统一用负极大搜索代替
int NegaMax(int depth){ int best = -INFINITY; if (depth == 0) { return evaluation(); } GenerateLegalMoves(); for (each move m) { do move m; val = - NegaMax(depth - 1); undo move m; if (val > best) { best = val; } } return best;}width="728" scrolling="no" height="90" frameborder="0" align="middle" marginwidth="0" marginheight="0" src="http://download1.csdn.net/down3/20070601/01184120111.htm">
扫一扫
3801
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
