Sudoku is one of the metaphysical techniques. If you understand the essence of it, you will have the feeling of being immortal, laughing and laughing with joy.............................................................(audience:"We need't such a geomancer!You can get out!!")

  Oh,I'm sorry.Then let's getting down to business as quickly as possible.

The meaning of the problem isn't hard for us to understand.And it is more friendly to those who have played sudoku.But don't worry if you didn't heard about this kind of game,for there's few skills of formal competition.What you have to do is to memorize its basic rules.

  OK.Now,let's think of the solution of this problem.You may want to enumerate all the alternative numbers in the blanks,then check if it's lawful.Good!it's the first algorithm we've thought of.But there's great room for progress.In fact,when people play sudoku,they will write down candidates in each blank.We can record which number we can use in one blank and try using it.To realize it,we can use state compression to save alternative numbers in each row,column and gong.After that,we can choose the blank which has the least candidates to have test-fill until all the blanks are filled.Don't forget to flash back after updating.

  Up to now,we are able to pass POJ2676,but there's still some distance from passing POJ3074.Think of how we search for the blank which has the least candidates and the use of lowbit.We can use an array to record the position of 1 in lowbit(x),and another one array to preprocess the number of 1 in each binary number.My code is below.

  After you get AC in these two sudokus,you can have a try in a harder promblem POJ3076.But you need a more skillful way.Can you find it?

 #include<bits/stdc++.h>

 using namespace std;

 int row[],column[],house[][],h[],cnt[];
char a[][],c;
bool flag; inline int getchoice(int x,int y){
return row[x]&column[y]&house[x/+][y/+];
} inline int lowbit(int x){
return x&-x;
} void dfs(int x,int y){
if(flag) return;
if(x==){
for(int i=;i<=;i++)
for(int j=;j<=;j++) cout<<a[i][j];
cout<<endl;
flag=true;
return;
}
int hx=x/+,hy=y/+;//判断(x,y)在哪一宫
int choice=getchoice(x,y);
while(choice){
int num=lowbit(choice);
choice-=num;
int row0=row[x],column0=column[y],house0=house[hx][hy];
row[x]&=~num;column[y]&=~num;house[hx][hy]&=~num;
a[x][y]=h[num]+'';
int minn=,xx=,yy;
for(int i=;i<=;i++)
for(int j=;j<=;j++){
if(a[i][j]!='.') continue;
int cal=cnt[getchoice(i,j)];
if(cal<minn){
xx=i;yy=j;
minn=cal;
}
}
dfs(xx,yy);
row[x]=row0;column[y]=column0;house[hx][hy]=house0;
}
a[x][y]='.';
} int main(){
for(int i=;i<=;i++) h[<<i]=i;//预处理lowbit(x)后x中1的位置(即2^n中n的值)
for(int i=;i<;i++)
for (int j=i;j;j-=lowbit(j))
cnt[i]++;//题解中的好方法,预处理每一个二进制数中1的数量;
cin>>c;
while(c!='e'){
for(int i=;i<=;i++)
for(int j=;j<=;j++)
row[i]=column[j]=house[i/+][j/+]=;
a[][]=c;
for(int j=;j<=;j++) scanf("%c",&a[][j]);
for(int i=;i<=;i++)
for(int j=;j<=;j++) scanf("%c",&a[i][j]);
for(int i=;i<=;i++)
for(int j=;j<=;j++)
if(a[i][j]!='.'){
int num=a[i][j]-'';
num=pow(,num);
row[i]&=~num;
column[j]&=~num;
house[i/+][j/+]&=~num;//预处理每行、每列、每宫可用的数字
}
flag=false;
int minn=,xx=,yy;
for(int i=;i<=;i++)
for(int j=;j<=;j++){
if(a[i][j]!='.') continue;
int cal=cnt[getchoice(i,j)];
if(cal<minn){
xx=i;yy=j;
minn=cal;
}
}
dfs(xx,yy);
cin>>c;
}
return ;
}

Sudoku(POJ2676/3074)的更多相关文章

  1. POJ 3074 Sudoku (Dancing Links)

    传送门:http://poj.org/problem?id=3074 DLX 数独的9*9的模板题. 具体建模详见下面这篇论文.其中9*9的数独怎么转化到精确覆盖问题,以及相关矩阵行列的定义都在下文中 ...

  2. POJ 2676 Sudoku(深搜)

    Sudoku Time Limit : 4000/2000ms (Java/Other)   Memory Limit : 131072/65536K (Java/Other) Total Submi ...

  3. HDU - 5547 Sudoku(数独搜索)

    Description Yi Sima was one of the best counselors of Cao Cao. He likes to play a funny game himself ...

  4. UESTC - 1222 Sudoku(深搜)

    Yi Sima was one of the best counselors of Cao Cao. He likes to play a funny game himself. It looks l ...

  5. HDU 3111 Sudoku(精确覆盖)

    数独问题,输入谜题,输出解 既然都把重复覆盖的给写成模板了,就顺便把精确覆盖的模板也写好看点吧...赤裸裸的精确覆盖啊~~~水一水~~~然后继续去搞有点难度的题了... #include <cs ...

  6. LeetCode 36 Valid Sudoku(合法的数独)

    题目链接: https://leetcode.com/problems/valid-sudoku/?tab=Description   给出一个二维数组,数组大小为数独的大小,即9*9  其中,未填入 ...

  7. Sudoku(简单DFS)

    题目链接:http://acm.hdu.edu.cn/showproblem.php?pid=5547 数据比较少,直接暴力DFS,检验成立情况即可 AC代码:但是不知道为什么用scanf,print ...

  8. 【POJ - 2676】Sudoku(数独 dfs+回溯)

    -->Sudoku 直接中文 Descriptions: Sudoku对数独非常感兴趣,今天他在书上看到了几道数独题: 给定一个由3*3的方块分割而成的9*9的表格(如图),其中一些表格填有1- ...

  9. sudoku 心得 视觉消除法(Visual Elimination)

    虽然我是程序员,但这里只介绍人类的思维方法. 这个方法我是从这里看到的: https://www.learn-sudoku.com/visual-elimination.html Most peopl ...

随机推荐

  1. 第一次C语言程序设计

    C语言程序实验报告 实验项目: 1.字符与ASCII码 2.运算符与表达式的应用 3.顺序结构应用程序 4.数学函数的算法描述 5.鸡兔同笼的算法描述 6.确定坐标的算法描述 姓名:张时锋 实验地点: ...

  2. .net session 使用误区

    1. this.Session["username"] = null  HttpSessionState 内部使用 NameObjectCollection 类型的集合对象来存储用 ...

  3. MongoDB集群的搭建

    一.环境准备 1.Centos7 2.mongodb3.4.10 3.三台机器IP分别是:192.168.1.100.192.168.1.135.192.168.1.136 二.mongdb数据库的安 ...

  4. python--第二十一/二天总结

    Python的WEB框架 Bottle Bottle是一个快速.简洁.轻量级的基于WSIG的微型Web框架,此框架只由一个 .py 文件,除了Python的标准库外,其不依赖任何其他模块. 1 2 3 ...

  5. Babel插件:@babel/plugin-transform-runtime

    一 概述 每个Babel编译后的脚本文件,都以导入的方式使用Babel的帮助函数,而不是每个文件都复制一份帮助函数的代码. 1 优点 (1)提高代码重用性,缩小编译后的代码体积. (2)防止污染全局作 ...

  6. [剑指Offer]29-顺时针打印矩阵-Java

    题目链接 https://www.nowcoder.com/practice/9b4c81a02cd34f76be2659fa0d54342a?tpId=13&tqId=11172&t ...

  7. iOS内置麦克风选择方法

    模式中的 voicechat用于VoIP是由系统进行默认选择的最适合的麦克风 模式中的AVAudioSessionModeVideoRecording默认选择上麦克风,离摄像头最近的那个,主要用于VO ...

  8. paloalto防火墙版本升级

    1.准备工作:此部分不影响生产环境,可直接操作. 1)备份: 2)下载OS HA情况下,在主机下载完成后,选择 Sync To Peer(同步到对端)同步到备机. 2.安装更新 1)在备机上选择安装 ...

  9. docker与虚拟机的区别

    Docker 是一个开源的应用容器引擎,让开发者可以打包他们的应用以及依赖包到一个可移植的容器中,然后发布到任何流行的 Linux 机器上,也可以实现虚拟化.容器是完全使用沙箱机制,相互之间不会有任何 ...

  10. Kudu基本操作及概念

    Kudu:    针对 Apache Hadoop 平台而开发的列式存储管理器. 使用场景:    适用于那些既有随机访问,也有批量数据扫描的复合场景.    高计算量的场景.    使用了高性能的存 ...