Fools and Foolproof Roads CodeForces - 362D(并查集+优先队列)

Fools and Foolproof Roads

CodeForces - 362D


You must have heard all about the Foolland on your Geography lessons. Specifically, you must know that federal structure of this country has been the same for many centuries. The country consists of n cities, some pairs of cities are connected by bidirectional roads, each road is described by its length li.

The fools lived in their land joyfully, but a recent revolution changed the king. Now the king is Vasily the Bear. Vasily divided the country cities into regions, so that any two cities of the same region have a path along the roads between them and any two cities of different regions don't have such path. Then Vasily decided to upgrade the road network and construct exactly p new roads in the country. Constructing a road goes like this:

  1. We choose a pair of distinct cities u, v that will be connected by a new road (at that, it is possible that there already is a road between these cities).
  2. We define the length of the new road: if cities u, v belong to distinct regions, then the length is calculated as min(109, S + 1) (S — the total length of all roads that exist in the linked regions), otherwise we assume that the length equals 1000.
  3. We build a road of the specified length between the chosen cities. If the new road connects two distinct regions, after construction of the road these regions are combined into one new region.

Vasily wants the road constructing process to result in the country that consists exactly of q regions. Your task is to come up with such road constructing plan for Vasily that it meets the requirement and minimizes the total length of the built roads.


Input

The first line contains four integers n (1 ≤ n ≤ 105), m (0 ≤ m ≤ 105), p (0 ≤ p ≤ 105), q (1 ≤ q ≤ n) — the number of cities in the Foolland, the number of existing roads, the number of roads that are planned to construct and the required number of regions.

Next m lines describe the roads that exist by the moment upgrading of the roads begun. Each of these lines contains three integers xi, yi, li: xi, yi — the numbers of the cities connected by this road (1 ≤ xi, yi ≤ n, xi ≠ yi), li — length of the road (1 ≤ li ≤ 109). Note that one pair of cities can be connected with multiple roads.

Output

If constructing the roads in the required way is impossible, print a single string "NO" (without the quotes). Otherwise, in the first line print word "YES" (without the quotes), and in the next p lines print the road construction plan. Each line of the plan must consist of two distinct integers, giving the numbers of the cities connected by a road. The road must occur in the plan in the order they need to be constructed. If there are multiple optimal solutions, you can print any of them.

Examples
Input
9 6 2 2
1 2 2
3 2 1
4 6 20
1 3 8
7 8 3
5 7 2
Output
YES
9 5
1 9
Input
2 0 1 2
Output
NO
Input
2 0 0 2
Output
YES
Note

Consider the first sample. Before the reform the Foolland consists of four regions. The first region includes cities 1, 2, 3, the second region has cities 4 and 6, the third region has cities 5, 7, 8, the fourth region has city 9. The total length of the roads in these cities is 11, 20, 5 and 0, correspondingly. According to the plan, we first build the road of length 6 between cities 5 and 9, then the road of length 23 between cities 1 and 9. Thus, the total length of the built roads equals 29.

题意:
给出一个无向图, 有n个点, m条边.
要求再添加p条边, 使得图中有q个连通块.
并使得所加边的权值和最小.
所加的边权值为min(1e9, S+1).
S是两个连通块的权值和(连通块的权值为所包含的边的权值和).

解题思路:
用并查集维护连通块, 用优先队列维护每个连通块的权值, 每次要减小连通块的数量(在两个连通块之间加边)的时候就从优先队列里面拿两个最小的连通块.

权值用long long啊!!!!

code:

#include <iostream>
#include <cstring>
#include <cstdio>
#include <algorithm>
#include <vector>
#include <queue>
using namespace std;
typedef long long ll;
const int maxn = 5e5+5;
const ll INF = 1e9;
struct node{
    int id;
    ll sum;
    bool operator < (const node& other)const{
        return sum > other.sum;
    }
};
int pre[maxn],num[maxn];
ll sum[maxn];
int vis[maxn];
vector<pair<int,int> >res;
priority_queue<node>que;

void init(int n){
    for(int i = 1; i <= n; i++){
        pre[i] = i;
        sum[i] = 0ll;
        num[i] = 1;
        vis[i] = 0;
    }
}
int Find(int x){
    if(x == pre[x])
        return x;
    else return pre[x] = Find(pre[x]);
}
void Union(int x,int y,ll w){
    int fx = Find(x),fy = Find(y);
    if(fx != fy){
        pre[fy] = fx;
        sum[fx] += sum[fy];
        num[fx] += num[fy];
    }
    sum[fx] += w;
    return;
}
int main(){
    int n,m,p,q;
    scanf("%d%d%d%d",&n,&m,&p,&q);
    init(n);
    while(m--){
        int x,y;
        ll w;
        scanf("%d%d%lld",&x,&y,&w);
        Union(x,y,w);
    }
    int region = 0;
    for(int i = 1; i <= n; i++){
        if(Find(i) == i)
            region++;//算算几个连通分量
    }
    region -= q;
    if(region < 0){
        printf("NO\n");
        return 0;
    }
    for(int i = 1; i <= n; i++){
        int fx = Find(i);
        if(!vis[fx]){//父亲节点入队,相当于把连通分量入队
            node tmp;
            tmp.id = fx;
            tmp.sum = sum[fx];
            que.push(tmp);
            vis[fx] = 1;
        }
    }
    while(!que.empty() && region--){
        node tmp1 = que.top();
        que.pop();
        if(que.empty()) break;
        node tmp2 = que.top();
        que.pop();
        res.push_back(make_pair(tmp1.id,tmp2.id));//连接两个连通分量存在答案数组中
        //连接后更新
        sum[tmp1.id] += (sum[tmp2.id] + min(INF,sum[tmp1.id]+sum[tmp2.id]+1ll));
        num[tmp1.id] += num[tmp2.id];
        pre[tmp2.id] = tmp1.id;
        tmp1.sum = sum[tmp1.id];
        que.push(tmp1);//更改后重新入队
        p--;
    }
    if(p < 0){//边不够
        printf("NO\n");
        return 0;
    }
    if(p > 0){
        int flag = 0;
        for(int i = 1; i <= n; i++){
            int fx = Find(i);
            if(num[fx] > 1){
                flag = fx;//记录下该连通分量的父亲
                break;
            }
        }
        if(!flag){//每个连通分量只有一个点,那么只能出现自连边但是题目不允许
            printf("NO\n");
            return 0;
        }
        int id = 0;
        for(int i = 1; i <= n; i++){
            if(Find(i) == flag && i != flag){
                id = i;
                break;//找到连通分量重除父亲外的另一个点
            }
        }
        while(p--){
            res.push_back(make_pair(flag,id));//一个连通分量中加边
        }
    }
    printf("YES\n");
    for(int i = 0; i < res.size(); i++){
        printf("%d %d\n",res[i].first,res[i].second);
    }
    return 0;
}

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。
内容概要:本文系统阐述了企业新闻发稿在生成式引擎优化(GEO)时代下的全渠道策略与效果评估体系,涵盖当前企业传播面临的预算、资源、内容与效果评估四大挑战,并深入分析2025年新闻发稿行业五大趋势,包括AI驱动的智能化转型、精准化传播、首发内容价值提升、内容资产化及数据可视化。文章重点解析央媒、地方官媒、综合门户和自媒体四类媒体资源的特性、传播优势与发稿策略,提出基于内容适配性、时间节奏、话题设计的策略制定方法,并构建涵盖品牌价值、销售转化与GEO优化的多维评估框架。此外,结合“传声港”工具实操指南,提供AI智能投放、效果监测、自媒体管理与舆情应对的全流程解决方案,并针对科技、消费、B2B、区域品牌四大行业推出定制化发稿方案。; 适合人群:企业市场/公关负责人、品牌传播管理者、数字营销从业者及中小企业决策者,具备一定媒体传播经验并希望提升发稿效率与ROI的专业人士。; 使用场景及目标:①制定科学的新闻发稿策略,实现从“流量思维”向“价值思维”转型;②构建央媒定调、门户扩散、自媒体互动的立体化传播矩阵;③利用AI工具实现精准投放与GEO优化,提升品牌在AI搜索中的权威性与可见性;④通过数据驱动评估体系量化品牌影响力与销售转化效果。; 阅读建议:建议结合文中提供的实操清单、案例分析与工具指南进行系统学习,重点关注媒体适配性策略与GEO评估指标,在实际发稿中分阶段试点“AI+全渠道”组合策略,并定期复盘优化,以实现品牌传播的长期复利效应。
评论
成就一亿技术人!
拼手气红包6.0元
还能输入1000个字符
 
红包 添加红包
表情包 插入表情
 条评论被折叠 查看
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