完美理论(最大权闭合子图)

Posted luoyibujue

tags:

篇首语:本文由小常识网(cha138.com)小编为大家整理,主要介绍了完美理论(最大权闭合子图)相关的知识,希望对你有一定的参考价值。

/*
最大权闭合子图模型 
枚举根, 然后选择包含根的连通块
那么就是选择儿子必须选择它的父亲
依赖关系就能够建立了
可以在这里提交 https://vijos.org/d/fastle/p/1011 
*/
#include<cstdio>
#include<algorithm>
#include<cstring>
#include<iostream>
#include<queue>
#define ll long long 
#define M 111
#define mmp make_pair
using namespace std;
int read()
{
    int nm = 0, f = 1;
    char c = getchar();
    for(; !isdigit(c); c = getchar()) if(c == '-') f = -1;
    for(; isdigit(c); c = getchar()) nm = nm * 10 + c - '0';
    return nm * f;
}
const int inf = 0x3e3e3e3e;
int head[M], to[M * M], nxt[M * M], a[M], cap[M * M], flow[M * M], deep[M], s, t, n, cnt;
vector<int> to1[M], to2[M];
void push(int vi, int vj, int fl)
{
//  cout << vi << " " << vj << " " << fl << "\n";
    cnt++, to[cnt] = vj, nxt[cnt] = head[vi], head[vi] = cnt, cap[cnt] = fl, flow[cnt] = 0;
    cnt++, to[cnt] = vi, nxt[cnt] = head[vj], head[vj] = cnt, cap[cnt] = 0, flow[cnt] = 0; 
}



void init()
{
    n = read(), s = n + 1, t = s + 1;
    for(int i = 1; i <= n; i++) vector<int>().swap(to1[i]), vector<int>().swap(to2[i]);
    for(int i = 1; i <= n; i++) a[i] = read();
    for(int i = 1; i < n; i++)
    {
        int vi = read(), vj = read();
        to1[vi].push_back(vj);
        to1[vj].push_back(vi);
    }
    for(int i = 1; i < n; i++)
    {
        int vi = read(), vj = read();
        to2[vi].push_back(vj);
        to2[vj].push_back(vi);
    }
}

bool bfs(int be, int ed)
{
    memset(deep, 0, sizeof(deep));
    deep[be] = 1;
    queue<int> q;
    q.push(be);
    while(!q.empty())
    {
        int now = q.front();
        q.pop();
        for(int i = head[now]; i; i = nxt[i])
        {
            int vj = to[i];
            if(deep[vj] || flow[i] >= cap[i]) continue;
            deep[vj] = deep[now] + 1;
            q.push(vj);
            if(vj == ed) return true;
        }
    }
    return false;
}

int dfs(int now, int ed, int fl)
{
    if(now == ed || fl == 0) return fl;
    int tot = 0, f;
    for(int i = head[now]; i; i = nxt[i])
    {
        int vj = to[i];
        if(deep[vj] != deep[now] + 1) continue;
        if(f = dfs(vj, ed, min(fl, cap[i] - flow[i])))
        {
            tot += f;
            flow[i] += f;
            flow[i ^ 1] -= f;
            fl -= f;
        }
        if(fl == 0) break;
    }
    if(tot == 0) deep[now] = -1;
    return tot;
}

void dfs1(int now, int fa)
{
    if(fa) push(now, fa, inf);
    for(int i = 0; i < to1[now].size(); i++)
    {
        int vj = to1[now][i];
        if(vj == fa) continue;
        dfs1(vj, now);
    }
}

void dfs2(int now, int fa)
{
    if(fa) push(now, fa, inf);
    for(int i = 0; i < to2[now].size(); i++)
    {
        int vj = to2[now][i];
        if(vj == fa) continue;
        dfs2(vj, now);
    }
}

int work(int now)
{
    cnt = 1;
    memset(head, 0, sizeof(head));
    dfs1(now, 0);
    dfs2(now, 0);
    int ans = 0;
    for(int i = 1; i <= n; i++) if(a[i] >= 0) push(s, i, a[i]), ans += a[i];
    else push(i, t, -a[i]);
    while(bfs(s, t)) ans -= dfs(s, t, inf);
    return ans;
}


int main()
{
    int T = read();
    while(T--)
    {
        init();
        int now = 0;
        for(int i = 1; i <= n; i++) now = max(now, work(i));
        cout << now << "\n";
    } 
    return 0;
}

以上是关于完美理论(最大权闭合子图)的主要内容,如果未能解决你的问题,请参考以下文章

最大权闭合子图(最小割)

最大权闭合子图 ( 最大流最小割模型 )

bzoj1497 最大获利(最大权闭合子图)

最大权闭合子图

最大权闭合子图进阶

最大权闭合子图