Emuskald considers himself a master of flow algorithms. Now he has completed his most ingenious program yet — it calculates the maximum flow in an undirected graph. The graph consists of n vertices and m edges. Vertices are numbered from 1 to n. Vertices 1 and n being the source and the sink respectively.

However, his max-flow algorithm seems to have a little flaw — it only finds the flow volume for each edge, but not its direction. Help him find for each edge the direction of the flow through this edges. Note, that the resulting flow should be correct maximum flow.

More formally. You are given an undirected graph. For each it's undirected edge (ai, bi) you are given the flow volume ci. You should direct all edges in such way that the following conditions hold:

  1. for each vertex v (1 < v < n), sum of ci of incoming edges is equal to the sum of ci of outcoming edges;
  2. vertex with number 1 has no incoming edges;
  3. the obtained directed graph does not have cycles.
 

Input

The first line of input contains two space-separated integers n and m (2 ≤ n ≤ 2·105, n - 1 ≤ m ≤ 2·105), the number of vertices and edges in the graph. The following m lines contain three space-separated integers ai, bi and ci (1 ≤ ai, bi ≤ n, ai ≠ bi, 1 ≤ ci ≤ 104), which means that there is an undirected edge from ai to bi with flow volume ci.

It is guaranteed that there are no two edges connecting the same vertices; the given graph is connected; a solution always exists.

 

Output

Output m lines, each containing one integer di, which should be 0 if the direction of the i-th edge is ai → bi (the flow goes from vertex ai to vertex bi) and should be 1 otherwise. The edges are numbered from 1 to m in the order they are given in the input.

If there are several solutions you can print any of them.

 

Sample Input

Input
3 3
3 2 10
1 2 10
3 1 5
Output
1
0
1
Input
4 5
1 2 10
1 3 10
2 3 5
4 2 15
3 4 5
Output
0
0
1
1
0
  可以发现这里有拓扑性质,可以直接做,O(N)复杂度。
 #include <iostream>
#include <cstring>
#include <cstdio>
#include <queue>
using namespace std;
const int N=,M=;
int cnt=,fir[N],nxt[M],to[M],cap[M];
int n,m,in[N],vis[N],ans[N];queue<int>q;
void addedge(int a,int b,int c){
nxt[++cnt]=fir[a];
to[fir[a]=cnt]=b;
cap[cnt]=c;
}
int main(){
scanf("%d%d",&n,&m);
for(int i=,a,b,c;i<=m;i++){
scanf("%d%d%d",&a,&b,&c);
addedge(a,b,c);addedge(b,a,c);
in[a]+=c;in[b]+=c;
}
for(int i=;i<n;i++)in[i]/=;
q.push();in[]=;vis[]=;
while(!q.empty()){
int x=q.front();q.pop();
for(int i=fir[x];i;i=nxt[i])
if(!vis[to[i]]){
in[to[i]]-=cap[i];
ans[i/]=i%;
if(in[to[i]]==){
q.push(to[i]);
vis[to[i]]=;
}
}
}
for(int i=;i<=m;i++)
printf("%d\n",ans[i]);
return ;
}

网络流相关(拓扑)CodeForces 269C:Flawed Flow的更多相关文章

  1. codeforces 269C Flawed Flow(网络流)

    Emuskald considers himself a master of flow algorithms. Now he has completed his most ingenious prog ...

  2. CodeForces - 269C Flawed Flow

    http://codeforces.com/problemset/problem/269/C 题目大意: 给定一个边没有定向的无法增广的残量网络且1是源点,n是汇点,给定每条边中的流.  让你把所有边 ...

  3. Codeforces 269C Flawed Flow (看题解)

    我好菜啊啊啊.. 循环以下操作 1.从队列中取出一个顶点, 把哪些没有用过的边全部用当前方向. 2.看有没有点的入度和 == 出度和, 如果有将当前的点加入队列. 现在有一个问题就是, 有没有可能队列 ...

  4. Codeforces 270E Flawed Flow 网络流问题

    题意:给出一些边,给出边的容量.让你为所有边确定一个方向使得流量最大. 题目不用求最大流, 而是求每条边的流向,这题是考察网络流的基本规律. 若某图有最大,则有与源点相连的边必然都是流出的,与汇点相连 ...

  5. [bzoj1565][NOI2009]植物大战僵尸_网络流_拓扑排序

    植物大战僵尸 bzoj1565 题目大意:给你一张网格图,上面种着一些植物.你从网格的最右侧开始进攻.每个植物可以对僵尸提供能量或者消耗僵尸的能量.每个植物可以保护一个特定网格内的植物,如果一个植物被 ...

  6. 网络流相关知识点以及题目//POJ1273 POJ 3436 POJ2112 POJ 1149

    首先来认识一下网络流中最大流的问题 给定一个有向图G=(V,E),把图中的边看做成管道,边权看做成每根管道能通过的最大流量(容量),给定源点s和汇点t,在源点有一个水源,在汇点有一个蓄水池,问s-t的 ...

  7. Oracle Spatial 中的弧段及弧相关拓扑错误

    1.报告说明 此报告用于验证下列问题: ORACLE SPATIAL 0.05m的最小拓扑容差值是否可以被修改 原始数据通过ARCGIS入库数据精度是否有损失 修改ORACLE SPATIAL图层的最 ...

  8. @codeforces - 708D@ Incorrect Flow

    目录 @description@ @solution@ @accepted code@ @details@ @description@ 给定一个有源点与汇点的图 G,并对于每一条边 (u, v) 给定 ...

  9. [模板] 网络流相关/最大流ISAP/费用流zkw

    最大流/ISAP 话说ISAP是真快...(大多数情况)吊打dinic,而且还好写... 大概思路就是: 在dinic的基础上, 动态修改层数, 如果终点层数 \(>\) 点数, break. ...

随机推荐

  1. WIN7 64位配置Oracle SQL Developer工具

    在使用Oracle SQL 的过程中,很多参考资料,辅导机构,各种书籍绝大多数都是使用PL/SQL进行讲解,但是问题是PL/SQL对WIN7 64位系统支持不好,网上有各种各样的配置教程,我尝试了很多 ...

  2. Codeforces 553C Love Triangles(图论)

    Solution: 比较好的图论的题. 要做这一题,首先要分析love关系和hate关系中,love关系具有传递性.更关键的一点,hate关系是不能成奇环的. 看到没有奇环很自然想到二分图的特性. 那 ...

  3. ACM YTU 挑战编程 字符串 Problem A: WERTYU

    Problem A: WERTYU Description A common typing error is to place yourhands on the keyboard one row to ...

  4. mongo数据库使用小结

    db.userId5555.aggregate({$unwind:"$tcjl"},{$match:{"_id":"0e549864-2a56-43c ...

  5. css sprites精灵技术:Html将所有图片放在一张图片上

    使用最近做的某项目常见页面作为联系素材: 分析:1.切图:步骤条可以切成四种图,即黄灰.红黄.红.灰. 2.html:需要五个li标签,每个包含一个图片及文字. 将要取得图片放到同一张图片上,从左到右 ...

  6. 如何更有效学习php开源项目的源码

    一.先把源代码安装起来,结合它的文档和手册,熟悉其功能和它的应用方式. 二.浏览源代码的目录结构,了解各个目录的功能. 三.经过以上两步后相信你对这个开源的产品有了一个初步的了解了,那现在就开始分析它 ...

  7. Python 手册——Python的非正式介绍

    在后面的例子中,区分输入和输出的方法是看是否有提示符(“>>> ”和“.. ”):想要重复这些例子的话,你就要在提示符显示后输入所有的一切:没有以提示符开始的行,是解释器输出的信息. ...

  8. VPN ,Bypass the FIrewall

    Bypass the China Firewall Methods November 16th, 2012Posted in , Tech With their assortment of techn ...

  9. uboot启动linux的过程

    一.概述 linux内核镜像常见到的有两种形式,zImage和uImage.这两种文件的格式稍有差别,所以启动这两种格式的内核镜像也会有所不同.目前,uboot只支持启动uImage类型的镜像,对zI ...

  10. 在IT学习中的“认识论”

    先有“感性认识”,再有“理性认识”.这句经典认识论,在IT领域应用广泛. 一个项目,首先是分析需求,也就是功能层面,这相当于先有“感性认识”:然后才是具体的技术实现,也就是“理性认识”. IT学习分析 ...