Coding Contest

Time Limit: 2000/1000 MS (Java/Others)    Memory Limit: 65536/65536 K (Java/Others)
Total Submission(s): 2653    Accepted Submission(s): 579

Problem Description

A coding contest will be held in this university, in a huge playground. The whole playground would be divided into N blocks, and there would be M directed paths linking these blocks. The i-th path goes from the ui-th block to the vi-th block. Your task is to solve the lunch issue. According to the arrangement, there are si competitors in the i-th block. Limited to the size of table, bi bags of lunch including breads, sausages and milk would be put in the i-th block. As a result, some competitors need to move to another block to access lunch. However, the playground is temporary, as a result there would be so many wires on the path.
For the i-th path, the wires have been stabilized at first and the first competitor who walker through it would not break the wires. Since then, however, when a person go through the i - th path, there is a chance of pi to touch
the wires and affect the whole networks. Moreover, to protect these wires, no more than ci competitors are allowed to walk through the i-th path.
Now you need to find a way for all competitors to get their lunch, and minimize the possibility of network crashing.

Input

The first line of input contains an integer t which is the number of test cases. Then t test cases follow.
For each test case, the first line consists of two integers N (N ≤ 100) and M (M ≤ 5000). Each of the next N lines contains two integers si and bi (si , bi ≤ 200).
Each of the next M lines contains three integers ui , vi and ci(ci ≤ 100) and a float-point number pi(0 < pi < 1).
It is guaranteed that there is at least one way to let every competitor has lunch.

Output

For each turn of each case, output the minimum possibility that the networks would break down. Round it to 2 digits.

Sample Input


1
4 4
2 0
0 3
3 0
0 3
1 2 5 0.5
3 2 5 0.5
1 4 5 0.5
3 4 5 0.5

Sample Output


0.50

思路:

这是一个最小费用最大流的变形题目。

要仔细读懂题目,题目要求的是崩溃的最小可能性。由于崩溃可能是由一条边或多条边一起的,所以正着求崩溃概率很难求,我们反着思考,用1去减每条边都不引起图崩溃的概率即为崩溃的概率。

由于求的是最小,而且边有容量限制,我们很自然的想到了最小费用最大流(对于不崩溃的概率应该是最大费用最大流)

建图时,某个点有人则想s连一条容量为人数的边,有食物则向t连一条容量为是无数的边,其他的边按照题目的数据相连即可,这个很好处理。

还有一个条件是某条路第一遍走的时候是没有花费的。我的处理方法是将一条容量为c,花费为w的边拆成两条边:容量为c-1,花费为w的一条边;容量为1,花费为1.0的一条边(花费为1.0是为了确保不会对最终结果产生影响)。

细节很多,由于概率要想乘所以初始化时要注意,算一条增广路径对总结果的贡献时也要注意是乘方而不是简单的相乘。

代码1:求增广路径时直接求最长路
  1 /*
2 * @FileName: D:\代码与算法\2017训练比赛\2016青岛区域赛\g.cpp
3 * @Author: Pic
4 * @Created Time: 2017/10/7 12:46:48
5 */
6 #include <bits/stdc++.h>
7 using namespace std;
8 const double INF = -1*10.0;
9 const int maxn=200+10;
10
11 struct Edge
12 {
13 int from,to,cap,flow;
14 double cost;
15 Edge(){}
16 Edge(int f,int t,int c,int fl,double co):from(f),to(t),cap(c),flow(fl),cost(co){}
17 };
18
19 struct MCMF
20 {
21 int n,m,s,t;
22 vector<Edge> edges;
23 vector<int> G[maxn];
24 bool inq[maxn]; //是否在队列
25 double d[maxn]; //Bellman_ford单源最短路径
26 int p[maxn]; //p[i]表从s到i的最小费用路径上的最后一条弧编号
27 int a[maxn]; //a[i]表示从s到i的最小残量
28
29 //初始化
30 void init(int n,int s,int t)
31 {
32 this->n=n, this->s=s, this->t=t;
33 edges.clear();
34 for(int i=0;i<n;++i) G[i].clear();
35 }
36
37 //添加一条有向边
38 void AddEdge(int from,int to,int cap,double cost)
39 {
40 edges.push_back(Edge(from,to,cap,0,cost));
41 edges.push_back(Edge(to,from,0,0,-cost));
42 m=edges.size();
43 G[from].push_back(m-2);
44 G[to].push_back(m-1);
45 }
46
47 //求一次增广路
48 bool BellmanFord(int &flow, double &cost)
49 {
50 for(int i=0;i<n;++i) d[i]=INF;
51 memset(inq,0,sizeof(inq));
52 d[s]=1.0, a[s]=1e9+30, inq[s]=true, p[s]=0;
53 queue<int> Q;
54 Q.push(s);
55 while(!Q.empty())
56 {
57 //cout<<"Q"<<endl;
58 int u=Q.front(); Q.pop();
59 inq[u]=false;
60 for(int i=0;i<G[u].size();++i)
61 {
62 Edge &e=edges[G[u][i]];
63 double tmp=(d[u]*e.cost);
64 //tmp=-1*fabs(tmp);
65 if( e.cap>e.flow && d[e.to] < tmp )
66 {
67 d[e.to]=tmp;
68 p[e.to]=G[u][i];
69 a[e.to]= min(a[u],e.cap-e.flow);
70 if(!inq[e.to]){ Q.push(e.to); inq[e.to]=true; }
71 }
72 }
73 }
74 //cout<<d[t]<<endl;
75 //if(fabs(d[t]-INF)<1e-3) return false;
76 if(d[t]<0) return false;
77 flow += a[t];
78 for(int i=0;i<a[t];i++) cost*=(d[t]);
79 int u=t;
80 while(u!=s)
81 {
82 edges[p[u]].flow += a[t];
83 edges[p[u]^1].flow -=a[t];
84 u = edges[p[u]].from;
85 }
86 return true;
87 }
88
89 //求出最小费用最大流
90 double Min_cost()
91 {
92 int flow=0;double cost=1.0;
93 while(BellmanFord(flow,cost)) ;
94 return cost;
95 }
96 }MM;
97 int main()
98 {
99 // freopen("data.in","r",stdin);
100 //freopen("data.out","w",stdout);
101 int t;
102 scanf("%d",&t);
103 while(t--){
104 int n,m;
105 scanf("%d%d",&n,&m);
106 MM.init(n+2,0,n+1);
107 int x,y;
108 for(int i=1;i<=n;i++){
109 scanf("%d%d",&x,&y);
110 if(x!=0)
111 MM.AddEdge(0,i,x,1.0);
112 if(y!=0)
113 MM.AddEdge(i,n+1,y,1.0);
114 }
115 int u,v,c;double w;
116 for(int i=1;i<=m;i++){
117 scanf("%d%d%d%lf",&u,&v,&c,&w);
118 //MM.AddEdge(u,v,c,w);
119 if(c>0){
120 MM.AddEdge(u,v,c-1,1.0-w);
121 MM.AddEdge(u,v,1,1.0);
122 }
123 }
124 double res=fabs(MM.Min_cost());
125 //cout<<res<<endl;
126 res=1.0-res;
127 if(res>=1.0) res=1.0;
128 printf("%.2lf\n",res);
129 }
130 return 0;
131 }

代码2:花费取负,求最小费用最大流,注意初始化的值

  1 /*
2 * @FileName: D:\代码与算法\2017训练比赛\2016青岛区域赛\g.cpp
3 * @Author: Pic
4 * @Created Time: 2017/10/7 12:46:48
5 */
6 #include <bits/stdc++.h>
7 using namespace std;
8 const double INF = 1e8;
9 const int maxn=200+10;
10
11 struct Edge
12 {
13 int from,to,cap,flow;
14 double cost;
15 Edge(){}
16 Edge(int f,int t,int c,int fl,double co):from(f),to(t),cap(c),flow(fl),cost(co){}
17 };
18
19 struct MCMF
20 {
21 int n,m,s,t;
22 vector<Edge> edges;
23 vector<int> G[maxn];
24 bool inq[maxn]; //是否在队列
25 double d[maxn]; //Bellman_ford单源最短路径
26 int p[maxn]; //p[i]表从s到i的最小费用路径上的最后一条弧编号
27 int a[maxn]; //a[i]表示从s到i的最小残量
28
29 //初始化
30 void init(int n,int s,int t)
31 {
32 this->n=n, this->s=s, this->t=t;
33 edges.clear();
34 for(int i=0;i<n;++i) G[i].clear();
35 }
36
37 //添加一条有向边
38 void AddEdge(int from,int to,int cap,double cost)
39 {
40 edges.push_back(Edge(from,to,cap,0,cost));
41 edges.push_back(Edge(to,from,0,0,-cost));
42 m=edges.size();
43 G[from].push_back(m-2);
44 G[to].push_back(m-1);
45 }
46
47 //求一次增广路
48 bool BellmanFord(int &flow, double &cost)
49 {
50 for(int i=0;i<n;++i) d[i]=INF;
51 memset(inq,0,sizeof(inq));
52 d[s]=1.0, a[s]=INF, inq[s]=true, p[s]=0;
53 queue<int> Q;
54 Q.push(s);
55 while(!Q.empty())
56 {
57 // cout<<"Q"<<endl;
58 int u=Q.front(); Q.pop();
59 inq[u]=false;
60 for(int i=0;i<G[u].size();++i)
61 {
62 Edge &e=edges[G[u][i]];
63 double tmp=(d[u]*e.cost);
64 tmp=-1*fabs(tmp);
65 if( e.cap>e.flow && d[e.to] > tmp )
66 {
67 d[e.to]=tmp;
68 p[e.to]=G[u][i];
69 a[e.to]= min(a[u],e.cap-e.flow);
70 if(!inq[e.to]){ Q.push(e.to); inq[e.to]=true; }
71 }
72 }
73 }
74 if(fabs(d[t]-INF)<1e3) return false;
75 flow += a[t];
76 for(int i=0;i<a[t];i++) cost*=(d[t]);
77 int u=t;
78 while(u!=s)
79 {
80 edges[p[u]].flow += a[t];
81 edges[p[u]^1].flow -=a[t];
82 u = edges[p[u]].from;
83 }
84 return true;
85 }
86
87 //求出最小费用最大流
88 double Min_cost()
89 {
90 int flow=0;double cost=1.0;
91 while(BellmanFord(flow,cost));
92 return cost;
93 }
94 }MM;
95 int main()
96 {
97 // freopen("data.in","r",stdin);
98 //freopen("data.out","w",stdout);
99 int t;
100 scanf("%d",&t);
101 while(t--){
102 int n,m;
103 scanf("%d%d",&n,&m);
104 MM.init(n+2,0,n+1);
105 int x,y;
106 for(int i=1;i<=n;i++){
107 scanf("%d%d",&x,&y);
108 if(x!=0)
109 MM.AddEdge(0,i,x,1.0);
110 if(y!=0)
111 MM.AddEdge(i,n+1,y,1.0);
112 }
113 int u,v,c;double w;
114 for(int i=1;i<=m;i++){
115 scanf("%d%d%d%lf",&u,&v,&c,&w);
116 //MM.AddEdge(u,v,c,w);
117 if(c>0){
118 MM.AddEdge(u,v,c-1,w-1.0);
119 MM.AddEdge(u,v,1,1.0);
120 }
121 }
122 double res=fabs(MM.Min_cost());
123 //cout<<res<<endl;
124 res=1.0-res;
125 if(res>=1.0) res=1.0;
126 printf("%.2lf\n",res);
127 }
128 return 0;
129 }


HDU–5988-Coding Contest(最小费用最大流变形)的更多相关文章

  1. HDU 5988 Coding Contest(最小费用最大流变形)

    Problem DescriptionA coding contest will be held in this university, in a huge playground. The whole ...

  2. HDU 5988.Coding Contest 最小费用最大流

    Coding Contest Time Limit: 2000/1000 MS (Java/Others)    Memory Limit: 65536/65536 K (Java/Others)To ...

  3. HDU5988/nowcoder 207G - Coding Contest - [最小费用最大流]

    题目链接:https://www.nowcoder.com/acm/contest/207/G 题目链接:http://acm.hdu.edu.cn/showproblem.php?pid=5988 ...

  4. hdu 1533 Going Home 最小费用最大流

    题目链接:http://acm.hdu.edu.cn/showproblem.php?pid=1533 On a grid map there are n little men and n house ...

  5. hdu 3667(拆边+最小费用最大流)

    题目链接:http://acm.hdu.edu.cn/showproblem.php?pid=3667 思路:由于花费的计算方法是a*x*x,因此必须拆边,使得最小费用流模板可用,即变成a*x的形式. ...

  6. hdu 3488(KM算法||最小费用最大流)

    Tour Time Limit: 3000/1000 MS (Java/Others)    Memory Limit: 65535/65535 K (Java/Others)Total Submis ...

  7. hdu 3395(KM算法||最小费用最大流(第二种超级巧妙))

    Special Fish Time Limit: 2000/1000 MS (Java/Others)    Memory Limit: 32768/32768 K (Java/Others)Tota ...

  8. hdu 1533 Going Home 最小费用最大流 入门题

    Going Home Time Limit: 10000/5000 MS (Java/Others)    Memory Limit: 65536/32768 K (Java/Others) Tota ...

  9. POJ 2195 & HDU 1533 Going Home(最小费用最大流)

    这就是一道最小费用最大流问题 最大流就体现到每一个'm'都能找到一个'H',但是要在这个基础上面加一个费用,按照题意费用就是(横坐标之差的绝对值加上纵坐标之差的绝对值) 然后最小费用最大流模板就是再用 ...

随机推荐

  1. bzoj2152 聪聪可可 (树形dp)

    大意: 给定树, 随机选两点, 求两点距离是3的倍数的概率. 树形dp入门水题, 枚举每个点作为lca时的答案即可. #include <iostream> #include <qu ...

  2. Centos 7.3 搭建php7,mysql5.7,nginx1.10.1,redis

    一.安装nginx 更新系统软件(非必要) # yum update 安装nginx 1.下载nginx # wget http://nginx.org/download/nginx-1.15.2.t ...

  3. 小知识:修改IDEA的模板

    小知识:修改IDEA的模板 有时候我们会发现,IDEA默认创建的模板并不是我们常用的.与其每次都在创建后进行修改,不如直接对模板进行修改. 给不知道怎么修改的同学指一下路: File->sett ...

  4. nexus 绑定负载均衡nginx反向代理后 遇到的https问题。

    1.今天搭建maven私服,下载安装好nexus运行后,通过IP可以直接访问,没有问题,如:http://123.123.123.123:8081 就可以进入主页面.没有任何问题. 2.但是他默认是h ...

  5. 基于Hadoop生态SparkStreaming的大数据实时流处理平台的搭建

    随着公司业务发展,对大数据的获取和实时处理的要求就会越来越高,日志处理.用户行为分析.场景业务分析等等,传统的写日志方式根本满足不了业务的实时处理需求,所以本人准备开始着手改造原系统中的数据处理方式, ...

  6. 关于小程序去除view/navigator 点击后默认阴影效果

    hover:class  :定义容器在被触发时的样式 通常无用,但若不去除则影响用户体验: 为避免被覆盖,约定在wxss底部添加class,比如: <!-- wxml --> <na ...

  7. 设置自己的bat运行文件-自己用,随时扩展

    start "" "E:\SEST_H5" start "" "C:\Program Files\Sublime Text 3\s ...

  8. SpringCloud之Eureka注册中心原理及其搭建

    一.Eureka简介 Eureka是Netflix开发的服务发现框架,本身是一个基于REST的服务,主要用于定位运行在AWS域中的中间层服务,以达到负载均衡和中间层服务故障转移的目的.SpringCl ...

  9. Centos使用光盘yum源

    yum查看所有源 yum repolist all 方法一:本机使用光盘源安装软件的设置 mkdir /media/cdrom mount /dev/cdrom  /media/cdrom vim / ...

  10. 7. Function Decorators and Closures

    A decorator is a callable that takes another function as argument (the decorated function). The deco ...