Codeforces Round #408 (Div. 2) D.Police Stations【Bfs+思维】

本文介绍了一种在保证每个普通节点到特殊节点距离不超过给定阈值的前提下,最大化裁减树形结构中边数量的算法。通过广度优先搜索策略确保了特殊节点覆盖的有效性和最优化,最终实现了道路维护成本的最小化。

D. Police Stations
time limit per test
2 seconds
memory limit per test
256 megabytes
input
standard input
output
standard output

Inzane finally found Zane with a lot of money to spare, so they together decided to establish a country of their own.

Ruling a country is not an easy job. Thieves and terrorists are always ready to ruin the country's peace. To fight back, Zane and Inzane have enacted a very effective law: from each city it must be possible to reach a police station by traveling at most d kilometers along the roads.

There are n cities in the country, numbered from 1 to n, connected only by exactly n - 1 roads. All roads are 1 kilometer long. It is initially possible to travel from a city to any other city using these roads. The country also has k police stations located in some cities. In particular, the city's structure satisfies the requirement enforced by the previously mentioned law. Also note that there can be multiple police stations in one city.

However, Zane feels like having as many as n - 1 roads is unnecessary. The country is having financial issues, so it wants to minimize the road maintenance cost by shutting down as many roads as possible.

Help Zane find the maximum number of roads that can be shut down without breaking the law. Also, help him determine such roads.

Input

The first line contains three integers n, k, and d (2 ≤ n ≤ 3·105, 1 ≤ k ≤ 3·105, 0 ≤ d ≤ n - 1) — the number of cities, the number of police stations, and the distance limitation in kilometers, respectively.

The second line contains k integers p1, p2, ..., pk (1 ≤ pi ≤ n) — each denoting the city each police station is located in.

The i-th of the following n - 1 lines contains two integers ui and vi (1 ≤ ui, vi ≤ n, ui ≠ vi) — the cities directly connected by the road with index i.

It is guaranteed that it is possible to travel from one city to any other city using only the roads. Also, it is possible from any city to reach a police station within d kilometers.

Output

In the first line, print one integer s that denotes the maximum number of roads that can be shut down.

In the second line, print s distinct integers, the indices of such roads, in any order.

If there are multiple answers, print any of them.

Examples
Input
6 2 4
1 6
1 2
2 3
3 4
4 5
5 6
Output
1
5
Input
6 3 2
1 5 6
1 2
1 3
1 4
1 5
5 6
Output
2
4 5 
Note

In the first sample, if you shut down road 5, all cities can still reach a police station within k = 4 kilometers.

In the second sample, although this is the only largest valid set of roads that can be shut down, you can print either 4 5 or 5 4 in the second line.


题目大意:

给你一个树,其中有N个点,N-1条无向边,现在有M个特殊点,问你最多可以切割多少条边,使得每个普通点到最近的特殊点的距离小于等于d.

保证有解。


思路:


保证有解,其实这个d就没有卵用了。

我们只要将这M个特殊点放到队列中,此时Bfs肯定有一个时间优先度。

那么我们直接Bfs.使得每次队头向下走一步,走到没有走过的点上边,对应记录每条边是否使用过即可。

这样去分块一定是最优的。

剩余没用过的边一定是最多的任意切割的边。


Ac代码:

#include<stdio.h>
#include<string.h>
#include<queue>
#include<vector>
using namespace std;
struct node
{
    int to,pos;
}now,nex;
vector<node >mp[300800];
int vis[300800];
int ans[300800];
int main()
{
    int n,m,k;
    while(~scanf("%d%d%d",&n,&m,&k))
    {
        queue<int >s;
        memset(vis,0,sizeof(vis));
        memset(ans,0,sizeof(ans));
        for(int i=1;i<=n;i++)mp[i].clear();
        for(int i=0;i<m;i++)
        {
            int x;
            scanf("%d",&x);
            s.push(x);
            vis[x]=1;
        }
        for(int i=1;i<=n-1;i++)
        {
            int x,y;
            scanf("%d%d",&x,&y);
            now.to=y;
            now.pos=i;
            mp[x].push_back(now);
            now.to=x;
            now.pos=i;
            mp[y].push_back(now);
        }
        while(!s.empty())
        {
            int u=s.front();
            s.pop();
            for(int i=0;i<mp[u].size();i++)
            {
                int v=mp[u][i].to;
                if(vis[v]==0)
                {
                    ans[mp[u][i].pos]=1;
                    vis[v]=1;
                    s.push(v);
                }
            }
        }
        int cnt=0;
        for(int i=1;i<=n-1;i++)if(ans[i]==0)cnt++;
        printf("%d\n",cnt);
        for(int i=1;i<=n-1;i++)
        {
            if(ans[i]==0)printf("%d ",i);
        }
        printf("\n");
    }
}













内容概要:本文介绍了ENVI Deep Learning V1.0的操作教程,重点讲解了如何利用ENVI软件进行深度学习模型的训练与应用,以实现遥感图像中特定目标(如集装箱)的自动提取。教程涵盖了从数据准备、标签图像创建、模型初始化与训练,到执行分类及结果优化的完整流程,并介绍了精度评价与通过ENVI Modeler实现一键化建模的方法。系统基于TensorFlow框架,采用ENVINet5(U-Net变体)架构,支持通过点、线、面ROI或分类图生成标签数据,适用于多/高光谱影像的单一类别特征提取。; 适合人群:具备遥感图像处理基础,熟悉ENVI软件操作,从事地理信息、测绘、环境监测等相关领域的技术人员或研究人员,尤其是希望将深度学习技术应用于遥感目标识别的初学者与实践者。; 使用场景及目标:①在遥感影像中自动识别和提取特定地物目标(如车辆、建筑、道路、集装箱等);②掌握ENVI环境下深度学习模型的训练流程与关键参数设置(如Patch Size、Epochs、Class Weight等);③通过模型调优与结果反馈提升分类精度,实现高效自动化信息提取。; 阅读建议:建议结合实际遥感项目边学边练,重点关注标签数据制作、模型参数配置与结果后处理环节,充分利用ENVI Modeler进行自动化建模与参数优化,同时注意软硬件环境(特别是NVIDIA GPU)的配置要求以保障训练效率。
评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值