在Java中向Negamax添加Alpha Beta修剪 [英] Adding Alpha Beta pruning to Negamax in Java

查看:101
本文介绍了在Java中向Negamax添加Alpha Beta修剪的处理方法,对大家解决问题具有一定的参考价值,需要的朋友们下面随着小编来一起学习吧!

问题描述

我正在用Java开发国际象棋游戏,并且(我认为)已经为AI玩家成功实现了Negamax。我在向其中添加alpha beta修剪以改进算法时遇到了一些麻烦。我已经尝试按照以下教程和示例代码进行操作,但是无法理解其工作原理。

I am making a chess game in Java and (I think) have successfully implemented Negamax for the AI player. I am having some trouble adding alpha beta pruning to this to improve the algorithm. I have tried following tutorials and example code but just can't get my head around how it works.

下面是我目前要采取最佳行动的代码:

Below is the code I currently have to get the best move:

private Move getBestMove() {
    System.out.println("Getting best move");
    System.out.println("Thinking...");

    List<Move> validMoves = generateMoves(true);
    int bestResult = Integer.MIN_VALUE;
    Move bestMove = null;

    for (Move move : validMoves) {

        executeMove(move);
        System.out.println("Evaluating: " + move);

        int evaluationResult = -evaluateNegaMax(this.lookForward, "", Integer.MIN_VALUE, Integer.MAX_VALUE);
        undoMove(move);

        if (evaluationResult > bestResult) {
            bestResult = evaluationResult;
            bestMove = move;
        }
    }
    System.out.println("Done thinking! The best move is: " + bestMove);
    return bestMove;
}

这是我尝试向我的(工作中)添加aplha-beta修剪的尝试negamax方法:

And here is my attempt at adding aplha-beta pruning to my (working) negamax method:

public int evaluateNegaMax(int lookForward, String indent, int alpha, int beta) {

    if (lookForward <= 0
            || this.chessGame.getGameState() == ChessGame.GAME_STATE_WHITE_WON
            || this.chessGame.getGameState() == ChessGame.GAME_STATE_BLACK_WON) {

        return evaluateState();
    }

    List<Move> moves = generateMoves(false);

    for (Move currentMove : moves) {

        System.out.println(indent + "Handling move: " + currentMove + " : " + alpha);
        if (currentMove == null) {
            continue;
        }

        executeMove(currentMove);

        alpha = Math.max(alpha, -evaluateNegaMax(lookForward-1, "    ", -beta, -alpha));

        if (alpha > beta) {
            break;
        }

        undoMove(currentMove);
    }
    return alpha;
}

最后是控制台的样子

Starting game flow
Looking 2 moves aheadExecuted: E/2 -> E/4
Tested 0 moves
Getting best move
Thinking...
Evaluating: B/8 -> A/6
Handling move: B/1 -> A/3 : -2147483648
    Handling move: A/8 -> B/8 : -2147483647
Handling move: B/1 -> C/3 : 2
    Handling move: B/8 -> A/8 : -2147483647
    Handling move: A/6 -> B/4 : -3
    Handling move: A/6 -> C/5 : -3
    Handling move: G/8 -> F/6 : -2
Handling move: D/1 -> E/2 : 2
    Handling move: B/8 -> A/8 : -2147483647
Handling move: D/1 -> F/3 : 2
    Handling move: A/8 -> B/8 : -2147483647
    Handling move: A/8 -> B/8 : -2147483647
    Handling move: F/6 -> E/4 : -32
    Handling move: F/6 -> G/4 : -17
    Handling move: F/6 -> D/5 : -17
Handling move: G/1 -> E/2 : 2
    Handling move: B/1 -> A/3 : -2147483647
    Handling move: B/1 -> C/3 : -29
    Handling move: E/1 -> F/1 : -28
    Handling move: E/2 -> G/1 : -19
    Handling move: E/2 -> C/3 : -19
    Handling move: E/2 -> G/3 : -19
    Handling move: E/2 -> D/4 : -19
Handling move: G/1 -> F/3 : 19
    Handling move: A/8 -> B/8 : -2147483647
Handling move: G/1 -> H/3 : 19
    Handling move: B/8 -> B/2 : -2147483647
Exception in thread "Thread-2" java.lang.NullPointerException
    at Chess.logic.ChessGame.movePiece(ChessGame.java:166)
    at Chess.ai.AiPlayerHandler.executeMove(AiPlayerHandler.java:158)
    at Chess.ai.AiPlayerHandler.evaluateNegaMax(AiPlayerHandler.java:84)
    at Chess.ai.AiPlayerHandler.getBestMove(AiPlayerHandler.java:47)
    at Chess.ai.AiPlayerHandler.getMove(AiPlayerHandler.java:31)
    at Chess.logic.ChessGame.waitForMove(ChessGame.java:125)
    at Chess.logic.ChessGame.startGame(ChessGame.java:95)
    at Chess.logic.ChessGame.run(ChessGame.java:338)
    at java.lang.Thread.run(Thread.java:745)

任何帮助将不胜感激。预先谢谢您。

Any help would be greatly appreciated. Thank you in advance.

推荐答案

我认为我正在使用它。如果以下问题的任何人都在等待响应,则代码如下:

I think I have it working. If anyone following this question was waiting for a response the code is as follows:

    public int evaluateNegaMax(int depth, String indent, int alpha, int beta) {
    if (depth <= 0
            || this.chessGame.getGameState() == ChessGame.GAME_STATE_WHITE_WON
            || this.chessGame.getGameState() == ChessGame.GAME_STATE_BLACK_WON) {

        return evaluateState();
    }

    List<Move> moves = generateMoves(false);
    int bestValue = Integer.MIN_VALUE;

    for (Move currentMove : moves) {

        executeMove(currentMove);
        int value = -evaluateNegaMax(depth - 1, indent + "    ", -beta, -alpha);
        System.out.println(indent + "Handling move: " + currentMove + " : " + value);
        undoMove(currentMove);
        counter++;

        if (value > bestValue) {
            bestValue = value;
        }

        if (bestValue > alpha) {
            alpha = bestValue;
        }

        if (bestValue >= beta) {
            break;
        }
    }
    System.out.println(indent + "max: " + alpha);
    return alpha;
}

这篇关于在Java中向Negamax添加Alpha Beta修剪的文章就介绍到这了,希望我们推荐的答案对大家有所帮助,也希望大家多多支持IT屋!

查看全文
登录 关闭
扫码关注1秒登录
发送“验证码”获取 | 15天全站免登陆