一款html5小游戏，主要功能在于js，带一套皮肤、内置ai算法，有能力的可以自行修改。
主要看下AI.js
贴出一部分代码学习下，
//人工智能初始化
AI.init = function(pace){
var bill = AI.historyBill || com.gambit; //开局库
if (bill.length){
var len=pace.length;
var arr=[];
//先搜索棋谱
for (var i=0;i< bill.length;i++){
if (bill[i].slice(0,len)==pace) {
arr.push(bill[i]);
}
}
if (arr.length){
var inx=Math.floor( Math.random() * arr.length );
AI.historyBill = arr ;
return arr[inx].slice(len,len+4).split("");
}else{
AI.historyBill = [] ;
}
}
//如果棋谱里面没有，人工智能开始运作
var initTime = new Date().getTime();
AI.treeDepth=play.depth;
//AI.treeDepth=4;
AI.number=0;
AI.setHistoryTable.lenght = 0
var val=AI.getAlphaBeta(-99999 ,99999, AI.treeDepth, com.arr2Clone(play.map),play.my);
//var val = AI.iterativeSearch(com.arr2Clone(play.map),play.my)
if (!val||val.value==-8888) {
AI.treeDepth=2;
val=AI.getAlphaBeta(-99999 ,99999, AI.treeDepth, com.arr2Clone(play.map),play.my);
}
//var val = AI.iterativeSearch(com.arr2Clone(play.map),play.my);
if (val&&val.value!=-8888) {
var man = play.mans[val.key];
var nowTime= new Date().getTime();
com.get("moveInfo").innerHTML='<h3>AI搜索结果：</h3>最佳着法：'+
com.createMove(com.arr2Clone(play.map),man.x,man.y,val.x,val.y)+
'<br />搜索深度：'+AI.treeDepth+'<br />搜索分支：'+
AI.number+'个 <br />最佳着法评估：'+
val.value+'分'+
' <br />搜索用时：'+
(nowTime-initTime)+'毫秒'
return [man.x,man.y,val.x,val.y]
}else {
return false;
}
}
//迭代加深搜索着法
AI.iterativeSearch = function (map, my){
var timeOut=100;
var initDepth = 1;
var maxDepth = 8;
AI.treeDepth=0;
var initTime = new Date().getTime();
var val = {};
for (var i=initDepth; i<=maxDepth; i++){
var nowTime= new Date().getTime();
AI.treeDepth=i;
AI.aotuDepth=i;
var val = AI.getAlphaBeta(-99999, 99999, AI.treeDepth , map ,my)
if (nowTime-initTime > timeOut){
return val;
}
}
return false;
}
//取得棋盘上所有棋子
AI.getMapAllMan = function (map, my){
var mans=[];
for (var i=0; i<map.length; i++){
for (var n=0; n<map[i].length; n++){
var key = map[i][n];
if (key && play.mans[key].my == my){
play.mans[key].x = n;
play.mans[key].y = i;
mans.push(play.mans[key])
}
}
}
return mans;
}
/*
//取得棋谱所有己方棋子的着法
AI.getMoves = function (map, my, txtMap){
var highMores = []; //优先级高的着法
var manArr = AI.getMapAllMan (map, my);
var moves = [];
var history=AI.historyTable[txtMap];
for (var i=0; i<manArr.length; i++){
var man = manArr[i];
var val=man.bl(map);
for (var n=0; n<val.length; n++){
if (history){
highMores.push([man.x,man.y,val[n][0],val[n][1],man.key])
}else{
moves.push([man.x,man.y,val[n][0],val[n][1],man.key])
}
}
}
return highMores.concat(moves);
}
*/
//取得棋谱所有己方棋子的着法
AI.getMoves = function (map, my){
var manArr = AI.getMapAllMan (map, my);
var moves = [];
var foul=play.isFoul;
for (var i=0; i<manArr.length; i++){
var man = manArr[i];
var val=man.bl(map);
for (var n=0; n<val.length; n++){
var x=man.x;
var y=man.y;
var newX=val[n][0];
var newY=val[n][1];
//如果不是长将着法
if (foul[0]!=x || foul[1]!=y || foul[2]!=newX || foul[3]!=newY ){
moves.push([x,y,newX,newY,man.key])
}
}
}
return moves;
}
//A:当前棋手value/B:对手value/depth：层级
AI.getAlphaBeta = function (A, B, depth, map ,my) {
//var txtMap= map.join();
//var history=AI.historyTable[txtMap];
// if (history && history.depth >= AI.treeDepth-depth+1){
// return history.value*my;
//}
if (depth == 0) {
return {"value":AI.evaluate(map , my)}; //局面评价函数;
}
var moves = AI.getMoves(map , my ); //生成全部走法;
//这里排序以后会增加效率
for (var i=0; i < moves.length; i++) {
//走这个走法;
var move= moves[i];
var key = move[4];
var oldX= move[0];
var oldY= move[1];
var newX= move[2];
var newY= move[3];
var clearKey = map[ newY ][ newX ]||"";
map[ newY ][ newX ] = key;
delete map[ oldY ][ oldX ];
play.mans[key].x = newX;
play.mans[key].y = newY;
if (clearKey=="j0"||clearKey=="J0") {//被吃老将,撤消这个走法;
play.mans[key] .x = oldX;
play.mans[key] .y = oldY;
map[ oldY ][ oldX ] = key;
delete map[ newY ][ newX ];
if (clearKey){
map[ newY ][ newX ] = clearKey;
// play.mans[ clearKey ].isShow = false;
}
return {"key":key,"x":newX,"y":newY,"value":8888};
//return rootKey;
}else {
var val = -AI.getAlphaBeta(-B, -A, depth - 1, map , -my).value;
//val = val || val.value;
//撤消这个走法;
play.mans[key] .x = oldX;
play.mans[key] .y = oldY;
map[ oldY ][ oldX ] = key;
delete map[ newY ][ newX ];
if (clearKey){
map[ newY ][ newX ] = clearKey;
//play.mans[ clearKey ].isShow = true;
}
if (val >= B) {
//将这个走法记录到历史表中;
//AI.setHistoryTable(txtMap,AI.treeDepth-depth+1,B,my);
return {"key":key,"x":newX,"y":newY,"value":B};
}
if (val > A) {
A = val; //设置最佳走法;
if (AI.treeDepth == depth) var rootKey={"key":key,"x":newX,"y":newY,"value":A};
}
}
}
//将这个走法记录到历史表中;
//AI.setHistoryTable(txtMap,AI.treeDepth-depth+1,A,my);
if (AI.treeDepth == depth) {//已经递归回根了
if (!rootKey){
//AI没有最佳走法，说明AI被将死了，返回false
return false;
}else{
//这个就是最佳走法;
return rootKey;
}
}
return {"key":key,"x":newX,"y":newY,"value":A};
}
//奖着法记录到历史表
AI.setHistoryTable = function (txtMap,depth,value,my){
AI.setHistoryTable.lenght ++;
AI.historyTable[txtMap] = {depth:depth,value:value}
}
//评估棋局 取得棋盘双方棋子价值差
评论0