围棋代码（围棋代码java）

本篇文章给大家谈谈围棋代码，以及围棋代码java对应的知识点，希望对各位有所帮助，不要忘了收藏本站喔。

本文目录一览：

• 1、如何用javascript实现围棋游戏
• 2、求围棋的VB源代码~
• 3、围棋java源代码
• 4、vb6.0编写围棋源代码
• 5、自学围棋的AlphaGo Zero，你也可以造一个
• 6、急求 用c++编写的人机对弈的围棋的程序代码，最好有注释！谢啦！！

如何用javascript实现围棋游戏

head

meta http-equiv="Content-Type" content="text/html;charset=utf-8" /

title禅棋传说/title

style type="text/css"

div { position: absolute; width: 23px; height: 23px; }

.B0 { background-image: url('B0.gif'); }

.B1 { background-image: url('B1.gif'); }

.B2 { background-image: url('B2.gif'); }

.B3 { background-image: url('B3.gif'); }

.B4 { background-image: url('B4.gif'); }

.B5 { background-image: url('B5.gif'); }

.B6 { background-image: url('B6.gif'); }

.B7 { background-image: url('B7.gif'); }

.B8 { background-image: url('B8.gif'); }

.BX { background-image: url('BX.gif'); }

.D0 { background-image: url('D0.gif'); }

.D1 { background-image: url('D1.gif'); }

.C0 { background-image: url('C0.gif'); }

.C1 { background-image: url('C1.gif'); }

/style

/head

body

script type="text/javascript"

//![CDATA[

Array.prototype.indexOf = function (item) //给数组扩展一个indexOf方法,用来检索是否为"已知元素"

{

for ( var i=0; ithis.length; i++)

if (this[i] == item)

return i;

return -1;

};

var Site = //定义一个棋位类

{

Create: function(x, y) //棋位类的构造函数

{

var me = document.createElement("div"); //建一个div对象，将其扩展并封装成棋位。

document.body.appendChild(me); //附加到DOM树，实现棋位的呈现。

me.x = x; //记录棋位的X坐标

me.y = y; //记录棋位的Y坐标

me.style.left = x * 23 + "px"; //设置棋位水平方向的绝对位置

me.style.top = y * 23 + "px"; //设置棋位垂直方向的绝对位置

// var s = ((x-9)%9?0:(x-9)/9)+1+(((y-9)%9?0:(y-9)/9)+1)*3; //计算背景式样,这个算法有冗余,简化为下面的算法.

var s = parseInt((x-9)/9)+1+(parseInt((y-9)/9)+1)*3; //计算背景式样

// me._backStyle = "B" + ((s==4(x/3)%2==1(y/3)%2==1) ? "X" : s);

me._backStyle = "B" + (((x/3)%2==1(y/3)%2==1) ? "X" : s);

me.Fill = this.Fill; //关联一个填充棋位的方法。

me.Tight = this.Tight; //关联计算紧气方法。

me.Kill = this.Kill; //关联计算死子方法。

me.onclick = this.Play; //绑定onclick事件到Play方法。

me.Fill(); //初始填充空子。

return me; //返回棋位对象，其实是一个封装了的div对象。

},

Fill: function(dot, going) //填充棋子的方法,going~~是否"当前步".

{

if ( dot == undefined )

this.className = this._backStyle //无子，就设置为背景式样。

else

this.className = (going ? "C" : "D") + dot; //有子,区别对待"当前步"

this.dot = dot; //保存棋子状态

},

Play: function() //行棋方法，由onclick事件触发

{

if ( this.dot == undefined ) //落子点必须为无子状态,否则不处理.

{

var deads = this.Kill(current^1); //计算可以杀死的子,current为当前颜色,则current^1为对方颜色

if (deads.length == 1 this == rob) return; //确认的打劫状态后,直接拒绝;如果此步落子可以提不止一子,则不是打劫

for(var i=0; ideads.length; i++)

deads[i].Fill(); //按照死子列清空,one by one

if(i==1)

rob = deads[0] //如果此步只提了一个子,记录被提位置为打劫位置.

else if (i0 || !this.Tight(current)) //这里有一个影响代码效率的判断,对于能产生提子的操做,不需要再判断是否"紧气禁入";此外,如果要引入"变穷为禁"的规则,这里需要改变算法

rob = null //清打劫位,这个条件是判断,如果能够提多子,或者不是禁止着手,则解除"打劫"状态

else return; //这里对应的返回,应该是只有无气禁入点的情况.

sound.play(); //落子有声！

var step = Tracks[Tracks.length-1];

if(step) step.site.Fill(step.site.dot); //更新此前一"步"的子的位图

this.Fill(current, true); //填入当前"步"的子

Tracks.push( new Step(this, deads) );

current ^= 1; //用1来异或，正好反转黑白棋子。

var disline = document.getElementById('list');

disline.value += ((current ? '黑：': '白：')+Tracks[Tracks.length-1].site.x+' '+Tracks[Tracks.length-1].site.y+'\n');

disline.scrollTop = disline.scrollHeight;

};

},

Tight: function (dot) //计算紧气的块,此时如果dot==undefined,则对应提子的遍历,this指向当前点.

{

var life = this.dot == undefined ? this : undefined; //life为"气"的定义;当前位无子则算一口气,对应落子的遍历,当前位置必然无子.

dot = dot == undefined ? this.dot : dot; //这个逻辑是对应提子的遍历

if (dot == undefined) return undefined;

var block = this.dot == undefined ? [] : [this]; //定义的"块",在提子的遍历中,发现子就放入块,再判断有多少"气".

var i = this.dot == undefined ? 0 : 1;

var site = this;

while (true)

{

for(var dx=-1;dx=1;dx++) for(var dy=-1;dy=1;dy++) if(!dx^!dy)

{

link = GetSite(site.x + dx, site.y + dy);

if (link) //判断目标位置的上下左右,有位则继续,"无位"对应棋盘以外的区域.

if (link.dot != undefined) //有子,则判断是否为同色,连"块"

{

if (link.dot == dot block.indexOf(link) 0 )

block.push(link); //此一段为"块"的遍历,条件是找四周的同色子,找到后判断是否为新"知道","新"则放入"块"

}

else if (!life) //无子,则更新"气"

life = link

else if (life != link)

return undefined; //在提子的遍历中,如果发现有两"气"了,则无须再算

};

if ( i = block.length) break; //"块"的遍历结束条件为,列表的最后一个对象没有产生新的相邻子.

site = block[i];

i ++;

};

return block; //返回只有一口气的块,在提子的遍历中,这对应了可能被提掉的子列,再去判断是否为"打劫"

},

Kill: function(dot) //计算杀死的子,目前this指向落子点,dot指向需要判断的死子颜色.

{

var deads = []; //定义死子列

for(var dx=-1;dx=1;dx++) for(var dy=-1;dy=1;dy++) if(!dx^!dy) //异或,有且只有一个为真,可对应到相邻子.

{

var site = GetSite(this.x + dx, this.y + dy);

if (site (site.dot == dot))

{

var block = site.Tight(); //分别对上下左右进行遍历,查找被紧气的"块",然后合并成为"死子列"

if (block) deads = deads.concat(block); //concat 合并操作,此操作并未识别"同项",仅以围棋规则判断,对"打劫没有影响".

//如果需要准确的提子数目统计,这里的算法需要更新.

};

};

return deads; //返回可以提子的死子块

}

}; //棋位类 Site

var Board = new Array(19); //全局的Board数组，表示棋盘。

var Tracks = []; //行棋线索数组，数组元素是Step对象。

var current = 0; //当前要下的子，0表示黑子，1表示白子，交替。

var rob = null; //如果有打劫时，记录打劫位置。

for(var x = 0 ; x 19; x++)

{

Board[x] = new Array(19);

for(var y = 0; y 19; y++)

Board[x][y] = Site.Create(x, y); //按位置创建棋位对象。

};

if (navigator.userAgent.indexOf(' MSIE ') -1) //为IE浏览器构造声音对象

{

var sound = document.body.appendChild(document.createElement("bgsound"));

sound.play = function(){this.src = "play.wav"};

}

else //为Firefox等其他浏览器构造声音对象

{

var sound = document.body.appendChild(document.createElement("span"));

sound.play = function(){this.innerHTML = "bgsound src='play.wav'"};

};

document.body.oncontextmenu = function() //右键驱动悔棋事件

{

var step = Tracks.pop();

if (step)

{

step.site.Fill();

for (var i=0; istep.deads.length; i++)

step.deads[i].Fill(current);

step = Tracks[Tracks.length-1];

if (step) step.site.Fill(current, true)

current ^= 1; //反转黑白棋子。

};

return false; //不弹出菜单。

};

function GetSite(x, y) //从棋盘取棋位的函数，越界不抛出异常。

{

if (x=0 x19 y=0 y19)

return Board[x][y];

};

function Step(site, deads) //棋步类，记录每一步棋的状态

{

this.site = site; //记录棋步的位置

this.deads = deads; //记录被当前棋步杀死的棋子集合

};

function PrintWay() //行棋路线

{

var str='', coler='';

for (var i=0;i Tracks.length;i++)

{

step = Tracks[i];

coler = (i%2) ? "白" : "黑";

str=str+"第"+(i+1)+"步"+coler+"方 X"+step.site.x+" Y"+step.site.y+" \n";

}

alert(str);

}

document.body.ondblclick = PrintWay;

document.onkeypress = function(event)

{

var k = (window.event ? window.event.keyCode : event.which) - 49; //按'1'可以进入自动摆棋操作,'1'=0x31=49

if(k0 || k1) return;

for(var x=0; x19; x++) for(var y=0; y19; y++) Board[x][y].Fill();

Tracks.length = 0;

current = 0;

with(goes[k]) for(var i=0; ilength;i+=3)

Board[charCodeAt(i+1)-65][charCodeAt(i)-65].Fill(charCodeAt(i+2)-48);

};

/body

/html

求围棋的VB源代码~

来~加 我 的QQ405557154 我做了一个，还有一同学做的，还有一是网上的朋友做的呵呵

Private Sub Timer1_Timer()

Dim ms As Boolean

Dim Info, temp As String

Dim p, p1, p2, i As Integer

Dim Ch

' Begin of Time Show Process

If ModemState LOGIN And SocketState CONNECTED Then

'it is not a multiusers game

Exit Sub

Else

If (S_R = 1) And (TURN = BLACKP) Then

Black_Time = Black_Time + Time - Start_Time

TimeB.Caption = CDate(Black_Time / 200)

Else

If (S_R = 1) And (TURN = WHITEP) Then

White_Time = White_Time + Time - Start_Time

TimeW.Caption = CDate(White_Time / 200)

Else

If (S_R = 0) And (TURN = WHITEP) Then

Black_Time = Black_Time + Time - Start_Time

TimeB.Caption = CDate(Black_Time / 200)

Else

If (S_R = 0) And (TURN = BLACKP) Then

White_Time = White_Time + Time - Start_Time

TimeW.Caption = CDate(White_Time / 200)

End If

End If

End If

End If

End If

'End If

' End of Time Show Process

'Begin of winsockt process

If SocketState = CONNECTED And Begin_Flag = 1 Then

ms = Net.Message_Exist

If ms = False Then

Exit Sub

End If

Info = Net.WaitForValue(Chr$(26), 5)

If g_ErrorCode = 1 Then

'Some error such as Timeout occured

Exit Sub

End If

p1 = InStr(Info, "B")

p2 = InStr(Info, "E|;")

If p1 = 0 Or p2 = 0 Then

Exit Sub

End If

temp = Mid$(Info, p1 + 1, p2 - p1 - 1)

ParseLine (temp)

Msg(Msg_No).No = CInt(ParseArray(1))

Msg(Msg_No).Color = CInt(ParseArray(2))

If IsNumeric(ParseArray(3)) Then

Msg(Msg_No).X = CInt(ParseArray(3))

Msg(Msg_No).Y = CInt(ParseArray(4))

Else

Msg(Msg_No).X = ParseArray(3)

Msg(Msg_No).Y = ParseArray(4)

End If

If Msg(Msg_No).Color = GIVEUP Then

Beep

MsgBox ("对方已经认输了")

Net.Winsock1.SendData ("R_O" + Chr$(26))

Pause 3

Call Begin_Click

Exit Sub

End If

If Side = BLACKP Then

p = Draw_Point(Msg(Msg_No).X, Msg(Msg_No).Y, WHITEP)

Record(Step).Color = WHITEP

step_show.Cls

step_show.Print Step

TURN = BLACKP

Else

p = Draw_Point(Msg(Msg_No).X, Msg(Msg_No).Y, BLACKP)

Record(Step).Color = BLACKP

step_show.Cls

step_show.Print Step

TURN = WHITEP

End If

Record(Step).X = Msg(Msg_No).X

Record(Step).Y = Msg(Msg_No).Y

Step = Step + 1

S_R = 1

R_R = 0

p = Count_All_Gas

If (Msg(Msg_No).X 0 And Msg(Msg_No).Y 0 _

And Msg(Msg_No).X 20 And Msg(Msg_No).Y 20) Then

Board(Msg(Msg_No).X, Msg(Msg_No).Y).Current = True

Refresh_Board

Board(Msg(Msg_No).X, Msg(Msg_No).Y).Current = False

End If

Msg_No = Msg_No + 1

Pause 1

Net.Winsock1.SendData ("R_O" + Chr$(26))

Exit Sub

End If

'End of process of winsocket

'Begin of modem process

If ModemState LOGIN Or R_R 1 Then

'It isn't a Inter_Modem Game

Exit Sub

End If

ms = Modem_F.Exist_Msg

If ms = False Then

'IO Port don't have any message

Exit Sub

End If

Info = Modem_F.WaitForValue(Chr$(26), 5)

'Wait a playing message

If g_ErrorCode = 1 Then

'Some error such as Timeout occured

Exit Sub

End If

p1 = InStr(Info, "B")

p2 = InStr(Info, "E|;")

If p1 = 0 Or p2 = 0 Then

Exit Sub

End If

temp = Mid$(Info, p1 + 1, p2 - p1 - 1)

ParseLine (temp)

Msg(Msg_No).No = CInt(ParseArray(1))

Msg(Msg_No).Color = CInt(ParseArray(2))

If IsNumeric(ParseArray(3)) Then

Msg(Msg_No).X = CInt(ParseArray(3))

Msg(Msg_No).Y = CInt(ParseArray(4))

Else

Msg(Msg_No).X = ParseArray(3)

Msg(Msg_No).Y = ParseArray(4)

End If

Modem_F.Comm1.InBufferCount = 0

'Clear Buffer

If Msg(Msg_No).Color = LOGOUT Then

Beep

MsgBox ("对方已经退出了")

Modem_F.Comm1.InBufferCount = 0

Modem_F.Comm1.Output = "R_O" + Chr$(26)

Cls

Step = 0

Start_Time = Time

Black_Time = 0

White_Time = 0

Exit Sub

End If

If Msg(Msg_No).Color = FINISHED Then

Beep

Ch = MsgBox("对方要求结束比赛，可以吗？", vbYesNo)

Modem_F.Comm1.InBufferCount = 0

If Ch = 6 Then

Modem_F.Comm1.Output = "YESR_O" + Chr$(26)

Else

Modem_F.Comm1.Output = "NOR_O" + Chr$(26)

Exit Sub

End If

PlayState = FINISHED

Count_Area.Enabled = True

End If

If Msg(Msg_No).Color = GIVEUP Then

Beep

MsgBox ("对方已经认输了")

Modem_F.Comm1.InBufferCount = 0

Modem_F.Comm1.Output = "R_O" + Chr$(26)

Call Begin_Click

Exit Sub

End If

If Msg(Msg_No).Color = TALK Then

Beep

MsgBox (Modem_F.His_Name.Text "说: " Msg(Msg_No).X)

Modem_F.Comm1.InBufferCount = 0

Modem_F.Comm1.Output = "R_O" + Chr$(26)

Exit Sub

End If

If Msg(Msg_No).Color = UNDO Then

Step = Step - 1

Beep

Modem_F.Comm1.InBufferCount = 0

Modem_F.Comm1.Output = "R_O" + Chr$(26)

Draw_Board

Ini_Board

For i = 1 To Step - 1

Board(Record(i).X, Record(i).Y).Current = False

p = Draw_Point(Record(i).X, Record(i).Y, Record(i).Color)

step_show.Cls

step_show.Print Step

p = Count_All_Gas

Next i

Board(Record(Step - 1).X, Record(Step - 1).Y).Current = True

Refresh_Board

Board(Record(Step - 1).X, Record(Step - 1).Y).Current = False

S_R = 1

R_R = 0

TURN = Side

Exit Sub

End If

Modem_F.Comm1.InBufferCount = 0

Modem_F.Comm1.Output = "R_O" + Chr$(26)

If Side = BLACKP Then

p = Draw_Point(Msg(Msg_No).X, Msg(Msg_No).Y, WHITEP)

Record(Step).Color = WHITEP

step_show.Cls

step_show.Print Step

Else

p = Draw_Point(Msg(Msg_No).X, Msg(Msg_No).Y, BLACKP)

Record(Step).Color = BLACKP

step_show.Cls

step_show.Print Step

End If

Record(Step).X = Msg(Msg_No).X

Record(Step).Y = Msg(Msg_No).Y

Step = Step + 1

S_R = 1

R_R = 0

p = Count_All_Gas

If (Msg(Msg_No).X 0 And Msg(Msg_No).Y 0 _

And Msg(Msg_No).X 20 And Msg(Msg_No).Y 20) Then

Board(Msg(Msg_No).X, Msg(Msg_No).Y).Current = True

Refresh_Board

Board(Msg(Msg_No).X, Msg(Msg_No).Y).Current = False

End If

Msg_No = Msg_No + 1

End Sub

Private Sub Timer2_Timer()

ShowS.Cls

ShowS.Print " :-):-):-):-):-) " + Show_String(C1) + " (:-(:-(:-(:-(:-"

C1 = C1 + 1

If C1 = 10 Then

C1 = 0

End If

Game_Time.Cls

Game_Time.Print " 现在时间: " Time

End Sub

Private Sub Set_Hand(h As Integer)

'设置让子（1-9）

Dim p As Integer

If h = 1 Then

'Not a Handicap game

Exit Sub

End If

部分代码。。

围棋java源代码

就给你一个吧，只不过没有添加棋子，你自己搞吧。import java.applet.*;

import java.awt.*;

import java.awt.event.*;

class ChessPad extends Panel implements MouseListener,ActionListener

{int x=-1,y=-1,棋子颜色=1;

Button button=new Button("重新开局");

TextField text_1=new TextField("请黑棋下子"),

text_2=new TextField();

ChessPad()

{setSize(440,440);

setLayout(null);setBackground(Color.pink);

addMouseListener(this);add(button);button.setBounds(10,5,60,26);

button.addActionListener(this);

add(text_1);text_1.setBounds(90,5,90,24);

add(text_2);text_2.setBounds(290,5,90,24);

text_1.setEditable(false);text_2.setEditable(false);

}

public void paint(Graphics g)

{for(int i=40;i=380;i+=20)

{g.drawLine(40,i,400,i);

}

g.drawLine(40,400,400,400);

for(int j=40;j=380;j+=20)

{g.drawLine(j,40,j,400);

}

g.drawLine(400,40,400,400);

g.fillOval(97,97,6,6);g.fillOval(337,97,6,6);

g.fillOval(97,337,6,6);g.fillOval(337,337,6,6);

g.fillOval(217,217,6,6);

}

public void mousePressed(MouseEvent e)

{if(e.getModifiers()==InputEvent.BUTTON1_MASK)

{x=(int)e.getX();y=(int)e.getY();

ChessPoint_black chesspoint_black=new ChessPoint_black(this);

ChessPoint_white chesspoint_white=new ChessPoint_white(this);

int a=(x+10)/2,b=(y+10)/2;

if(x/202||y/202||x/2019||y/2019)

{}

else

{

if(棋子颜色==1)

{this.add(chesspoint_black);

chesspoint_black.setBounds(a*20-7,b*20-7,16,16);

棋子颜色=棋子颜色*(-1);

text_2.setText("请白棋下子");

text_1.setText("");

}

else if(棋子颜色==-1)

{this.add(chesspoint_white);

chesspoint_white.setBounds(a*20-7,b*20-7,16,16);

棋子颜色=棋子颜色*(-1);

text_1.setText("请黑棋下子");

text_2.setText("");

}

}

}

}

public void mouseReleased(MouseEvent e){}

public void mouseEntered(MouseEvent e){}

public void mouseExited(MouseEvent e){}

public void mouseClicked(MouseEvent e){}

public void actionPerformed(ActionEvent e)

{this.removeAll();棋子颜色=1;

add(button);button.setBounds(10,5,60,26);

add(text_1);text_1.setBounds(90,5,90,24);

text_2.setText("");text_1.setText("请黑棋下子");

add(text_2);text_2.setBounds(290,5,90,24);

}

}

class ChessPoint_black extends Canvas implements MouseListener

{ChessPad chesspad=null;

ChessPoint_black(ChessPad p)

{setSize(20,20);chesspad=p;addMouseListener(this);

}

public void paint(Graphics g)

{g.setColor(Color.black);g.fillOval(0,0,14,14);

}

public void mousePressed(MouseEvent e)

{if(e.getModifiers()==InputEvent.BUTTON3_MASK)

{chesspad.remove(this);

chesspad.棋子颜色=1;

chesspad.text_2.setText("");chesspad.text_1.setText("请黑棋下子");

}

}

public void mouseReleased(MouseEvent e){}

public void mouseEntered(MouseEvent e){}

public void mouseExited(MouseEvent e){}

public void mouseClicked(MouseEvent e)

{if(e.getClickCount()=2)

chesspad.remove(this);

}

}

class ChessPoint_white extends Canvas implements MouseListener

{ChessPad chesspad=null;

ChessPoint_white(ChessPad p)

{setSize(20,20);addMouseListener(this);

chesspad=p;

}

public void paint(Graphics g)

{g.setColor(Color.white);g.fillOval(0,0,14,14);

}

public void mousePressed(MouseEvent e)

{if(e.getModifiers()==InputEvent.BUTTON3_MASK)

{chesspad.remove(this);

chesspad.棋子颜色=-1;

chesspad.text_2.setText("请白棋下子");chesspad.text_1.setText("");

}

}

public void mouseReleased(MouseEvent e){}

public void mouseEntered(MouseEvent e){}

public void mouseExited(MouseEvent e){}

public void mouseClicked(MouseEvent e)

{if(e.getClickCount()=2)

chesspad.remove(this);

}

}

public class Chess extends Frame

{ChessPad chesspad=new ChessPad();

Chess()

{setVisible(true);

setLayout(null);

Label label=new Label("单击左键下子，双击吃子，右击棋子悔棋",Label.CENTER);

add(label);label.setBounds(70,55,440,26);

label.setBackground(Color.orange);

add(chesspad);chesspad.setBounds(70,90,440,440);

addWindowListener(new WindowAdapter()

{public void windowClosing(WindowEvent e)

{System.exit(0);

}

});

pack();setSize(600,550);

}

public static void main (String args[])

{Chess chess=new Chess();

}

}

vb6.0编写围棋源代码

弈城的形势判断可以做评估函数

另,最好使用分割法,将棋盘分割,他走那你选择该分割棋盘内最好的应手(递归)或其他区块最好应手(这个可以先计算好再保存)

自学围棋的AlphaGo Zero，你也可以造一个

01

遥想当年，AlphaGo的Master版本，在完胜柯洁九段之后不久，就被后辈AlphaGo Zero (简称狗零) 击溃了。

从一只完全不懂围棋的AI，到打败Master，狗零只用了21天。

而且，它不需要用人类知识来喂养，成为顶尖棋手全靠自学。

如果能培育这样一只AI，即便自己不会下棋，也可以很骄傲吧。

于是，来自巴黎的少年Dylan Djian (简称小笛) ，就照着狗零的论文去实现了一下。

他给自己的AI棋手起名SuperGo，也提供了代码 (传送门见文底) 。

除此之外，还有教程——

一个身子两个头

智能体分成三个部分：

一是特征提取器 (Feature Extractor) ，二是策略网络 (Policy Network) ，三是价值网络(Value Network) 。

于是，狗零也被亲切地称为“双头怪”。特征提取器是身子，其他两个网络是脑子。

特征提取器

特征提取模型，是个残差网络 (ResNet) ，就是给普通CNN加上了跳层连接 (Skip Connection) ， 让梯度的传播更加通畅。

跳跃的样子，写成代码就是：

1class BasicBlock(nn.Module):

2 """

3 Basic residual block with 2 convolutions and a skip connection

4 before the last ReLU activation.

5 """

6

7 def __init__(self, inplanes, planes, stride=1, downsample=None):

8 super(BasicBlock, self).__init__()

9

10 self.conv1 = nn.Conv2d(inplanes, planes, kernel_size=3,

11 stride=stride, padding=1, bias=False)

12 self.bn1 = nn.BatchNorm2d(planes)

13

14 self.conv2 = nn.Conv2d(planes, planes, kernel_size=3,

15 stride=stride, padding=1, bias=False)

16 self.bn2 = nn.BatchNorm2d(planes)

17

18

19 def forward(self, x):

20 residual = x

21

22 out = self.conv1(x)

23 out = F.relu(self.bn1(out))

24

25 out = self.conv2(out)

26 out = self.bn2(out)

27

28 out += residual

29 out = F.relu(out)

30

31 return out

然后，把它加到特征提取模型里面去：

1class Extractor(nn.Module):

2 def __init__(self, inplanes, outplanes):

3 super(Extractor, self).__init__()

4 self.conv1 = nn.Conv2d(inplanes, outplanes, stride=1,

5 kernel_size=3, padding=1, bias=False)

6 self.bn1 = nn.BatchNorm2d(outplanes)

7

8 for block in range(BLOCKS):

9 setattr(self, "res{}".format(block), \

10 BasicBlock(outplanes, outplanes))

11

12

13 def forward(self, x):

14 x = F.relu(self.bn1(self.conv1(x)))

15 for block in range(BLOCKS - 1):

16 x = getattr(self, "res{}".format(block))(x)

17

18 feature_maps = getattr(self, "res{}".format(BLOCKS - 1))(x)

19 return feature_maps

策略网络

策略网络就是普通的CNN了，里面有个批量标准化 (Batch Normalization) ，还有一个全连接层，输出概率分布。

1class PolicyNet(nn.Module):

2 def __init__(self, inplanes, outplanes):

3 super(PolicyNet, self).__init__()

4 self.outplanes = outplanes

5 self.conv = nn.Conv2d(inplanes, 1, kernel_size=1)

6 self.bn = nn.BatchNorm2d(1)

7 self.logsoftmax = nn.LogSoftmax(dim=1)

8 self.fc = nn.Linear(outplanes - 1, outplanes)

9

10

11 def forward(self, x):

12 x = F.relu(self.bn(self.conv(x)))

13 x = x.view(-1, self.outplanes - 1)

14 x = self.fc(x)

15 probas = self.logsoftmax(x).exp()

16

17 return probas

价值网络

这个网络稍微复杂一点。除了标配之外，还要再多加一个全连接层。最后，用双曲正切 (Hyperbolic Tangent) 算出 (-1,1) 之间的数值，来表示当前状态下的赢面多大。

代码长这样——

1class ValueNet(nn.Module):

2 def __init__(self, inplanes, outplanes):

3 super(ValueNet, self).__init__()

4 self.outplanes = outplanes

5 self.conv = nn.Conv2d(inplanes, 1, kernel_size=1)

6 self.bn = nn.BatchNorm2d(1)

7 self.fc1 = nn.Linear(outplanes - 1, 256)

8 self.fc2 = nn.Linear(256, 1)

9

10

11 def forward(self, x):

12 x = F.relu(self.bn(self.conv(x)))

13 x = x.view(-1, self.outplanes - 1)

14 x = F.relu(self.fc1(x))

15 winning = F.tanh(self.fc2(x))

16 return winning

未雨绸缪的树

狗零，还有一个很重要的组成部分，就是蒙特卡洛树搜索 (MCTS) 。

它可以让AI棋手提前找出，胜率最高的落子点。

在模拟器里，模拟对方的下一手，以及再下一手，给出应对之策，所以提前的远不止是一步。

节点 (Node)

树上的每一个节点，都代表一种不同的局势，有不同的统计数据：

每个节点被经过的次数n，总动作值w，经过这一点的先验概率p，平均动作值q (q=w/n) ，还有从别处来到这个节点走的那一步，以及从这个节点出发、所有可能的下一步。

1class Node:

2 def __init__(self, parent=None, proba=None, move=None):

3 self.p = proba

4 self.n = 0

5 self.w = 0

6 self.q = 0

7 self.children = []

8 self.parent = parent

9 self.move = move

部署 (Rollout)

第一步是PUCT (多项式上置信树) 算法，选择能让PUCT函数 (下图) 的某个变体 (Variant) 最大化，的走法。

 写成代码的话——

1def select(nodes, c_puct=C_PUCT):

2 " Optimized version of the selection based of the PUCT formula "

3

4 total_count = 0

5 for i in range(nodes.shape[0]):

6 total_count += nodes[i][1]

7

8 action_scores = np.zeros(nodes.shape[0])

9 for i in range(nodes.shape[0]):

10 action_scores[i] = nodes[i][0] + c_puct * nodes[i][2] * \

11 (np.sqrt(total_count) / (1 + nodes[i][1]))

12

13 equals = np.where(action_scores == np.max(action_scores))[0]

14 if equals.shape[0] 0:

15 return np.random.choice(equals)

16 return equals[0]

结束 (Ending)

选择在不停地进行，直至到达一个叶节点 (Leaf Node) ，而这个节点还没有往下生枝。

1def is_leaf(self):

2 """ Check whether a node is a leaf or not """

3

4 return len(self.children) == 0

到了叶节点，那里的一个随机状态就会被评估，得出所有“下一步”的概率。

所有被禁的落子点，概率会变成零，然后重新把总概率归为1。

然后，这个叶节点就会生出枝节 (都是可以落子的位置，概率不为零的那些) 。代码如下——

1def expand(self, probas):

2 self.children = [Node(parent=self, move=idx, proba=probas[idx]) \

3 for idx in range(probas.shape[0]) if probas[idx] 0]

更新一下

枝节生好之后，这个叶节点和它的妈妈们，身上的统计数据都会更新，用的是下面这两串代码。

1def update(self, v):

2 """ Update the node statistics after a rollout """

3

4 self.w = self.w + v

5 self.q = self.w / self.n if self.n 0 else 0

1while current_node.parent:

2 current_node.update(v)

3 current_node = current_node.parent

选择落子点

模拟器搭好了，每个可能的“下一步”，都有了自己的统计数据。

按照这些数据，算法会选择其中一步，真要落子的地方。

选择有两种，一就是选择被模拟的次数最多的点。试用于测试和实战。

另外一种，随机 (Stochastically) 选择，把节点被经过的次数转换成概率分布，用的是以下代码——

1 total = np.sum(action_scores)

2 probas = action_scores / total

3 move = np.random.choice(action_scores.shape[0], p=probas)

后者适用于训练，让AlphaGo探索更多可能的选择。

三位一体的修炼

狗零的修炼分为三个过程，是异步的。

一是自对弈 (Self-Play) ，用来生成数据。

1def self_play():

2 while True:

3 new_player, checkpoint = load_player()

4 if new_player:

5 player = new_player

6

7 ## Create the self-play match queue of processes

8 results = create_matches(player, cores=PARALLEL_SELF_PLAY,

9 match_number=SELF_PLAY_MATCH)

10 for _ in range(SELF_PLAY_MATCH):

11 result = results.get()

12 db.insert({

13 "game": result,

14 "id": game_id

15 })

16 game_id += 1

二是训练 (Training) ，拿新鲜生成的数据，来改进当前的神经网络。

1def train():

2 criterion = AlphaLoss()

3 dataset = SelfPlayDataset()

4 player, checkpoint = load_player(current_time, loaded_version)

5 optimizer = create_optimizer(player, lr,

6 param=checkpoint['optimizer'])

7 best_player = deepcopy(player)

8 dataloader = DataLoader(dataset, collate_fn=collate_fn, \

9 batch_size=BATCH_SIZE, shuffle=True)

10

11 while True:

12 for batch_idx, (state, move, winner) in enumerate(dataloader):

13

14 ## Evaluate a copy of the current network

15 if total_ite % TRAIN_STEPS == 0:

16 pending_player = deepcopy(player)

17 result = evaluate(pending_player, best_player)

18

19 if result:

20 best_player = pending_player

21

22 example = {

23 'state': state,

24 'winner': winner,

25 'move' : move

26 }

27 optimizer.zero_grad()

28 winner, probas = pending_player.predict(example['state'])

29

30 loss = criterion(winner, example['winner'], \

31 probas, example['move'])

32 loss.backward()

33 optimizer.step()

34

35 ## Fetch new games

36 if total_ite % REFRESH_TICK == 0:

37 last_id = fetch_new_games(collection, dataset, last_id)

训练用的损失函数表示如下：

1class AlphaLoss(torch.nn.Module):

2 def __init__(self):

3 super(AlphaLoss, self).__init__()

4

5 def forward(self, pred_winner, winner, pred_probas, probas):

6 value_error = (winner - pred_winner) ** 2

7 policy_error = torch.sum((-probas *

8 (1e-6 + pred_probas).log()), 1)

9 total_error = (value_error.view(-1) + policy_error).mean()

10 return total_error

三是评估 (Evaluation) ，看训练过的智能体，比起正在生成数据的智能体，是不是更优秀了 (最优秀者回到第一步，继续生成数据) 。

1def evaluate(player, new_player):

2 results = play(player, opponent=new_player)

3 black_wins = 0

4 white_wins = 0

5

6 for result in results:

7 if result[0] == 1:

8 white_wins += 1

9 elif result[0] == 0:

10 black_wins += 1

11

12 ## Check if the trained player (black) is better than

13 ## the current best player depending on the threshold

14 if black_wins = EVAL_THRESH * len(results):

15 return True

16 return False

第三部分很重要，要不断选出最优的网络，来不断生成高质量的数据，才能提升AI的棋艺。

三个环节周而复始，才能养成强大的棋手。

有志于AI围棋的各位，也可以试一试这个PyTorch实现。

本来摘自量子位，原作 Dylan Djian。

代码实现传送门：

网页链接

教程原文传送门：

网页链接

AlphaGo Zero论文传送门：

网页链接

急求 用c++编写的人机对弈的围棋的程序代码，最好有注释！谢啦！！

int Count = 600;

AnsiString CaTimer(int Value)

{

AnsiString ss;

ss.printf("%02d:%02d:%02d",Value/(60*60), (Value%(60*60))/60,Value%60);

return ss;

}

//---------------------------------------------------------------------------

void __fastcall TForm1::Timer1Timer(TObject *Sender)

{

Count--;

Label1-Caption = CaTimer(Count) ;

}

// 要设置计时器 1秒钟 一次。 Timer1-Interval = 1000; (毫秒)

// 打开关闭 可以设置 Timer1 的属性 Enabled 为true , false

关于围棋代码和围棋代码java的介绍到此就结束了，不知道你从中找到你需要的信息了吗 ？如果你还想了解更多这方面的信息，记得收藏关注本站。