BACKUP - Backup Files

no tags 

You run an IT company that backs up computer data for large offices. Backing up data is not fun, and so you design your system so that the different offices can back up each others' data while you sit at home and play computer games instead.

The offices are all situated along a single street. You decide to pair up the offices, and for each pair of offices you run a network cable between the two buildings so that they can back up each others' data.

However, network cables are expensive. Your local telecommunications company will only give you k network cables, which means you can only arrange backups for k pairs of offices (2k offices in total). No office may belong to more than one pair (that is, these 2k offices must all be different). Furthermore, the telecommunications company charges by the kilometre. This means that you need to choose these k pairs of offices so that you use as little cable as possible. In other words, you need to choose the pairs so that, when the distances between the two offices in each pair are added together, the total distance is as small as possible.

As an example, suppose you had five clients with offices on a street as illustrated below. These offices are situated 1 km, 3 km, 4 km, 6km and 12km from the beginning of the street. The telecommunications company will only provide you with k = 2 cables.

The best pairing in this example is created by linking the first and second offices together, and linking the third and fourth offices together. This uses k = 2 cables as required, where the first cable has length 3km - 1km = 2 km, and the second cable has length 6km - 4km = 2 km. This pairing requires a total of 4km of network cables, which is the smallest total possible.

Input

Multiple test cases, the number of them will be given at the very first line.

For each test case:

The first line of input will contain the integers n and k, representing the number of offices on the street (2 <= n <= 100 000) and the number of available network cables (1 <= k <= n/2).

The following n lines will each contain a single integer (0 <= s <= 1 000 000 000), representing the distance of each office from the beginning of the street. These integers will appear in sorted order from smallest to largest. No two offices will share the same location.

Output

Output should consist of a single positive integer, giving the smallest total length of network cable required to join 2k distinct offices into k pairs.

Example

Input:
1
5 2
1
3
4
6
12 Output:
4 Explanation

The sample input above represents the example scenario described earlier.

Warning: large input/output data,be careful with certain languages

Blue Mary's Note: test data has been modified on Dec. 5, 2007. All the solutions have been rejudged.

  这题有个性质,就是你要匹配的任意两对点的连线不能重合, 否则就不是最优的了。利用这个性质去贪心解题。

 #include <cstdio>
#include <algorithm>
#include <cstring>
#include <queue>
#include <vector>
using std::priority_queue;
using std::vector;
typedef long long LL;
struct A{
int l, r, pos;
LL dis;
};
const int Maxn = ;
int s[Maxn], mark[Maxn], tot;
A l[Maxn];
struct cmp{
bool operator ()(A a,A b){
return a.dis > b.dis;
}
};
int main(){
#ifndef ONl_JUDGE
freopen ("backup.in", "r", stdin);
freopen ("backup.out", "w", stdout);
#endif
int n, k;
scanf ("%d%d", &n, &k);
priority_queue<A, vector<A>, cmp> q;
for (int i = ; i <= n; ++i)
scanf ("%d", s+i);
for(int i=;i<n;i++)
{
l[i].dis = s[i+]-s[i];
l[i].pos = i;
l[i].l = i-;
l[i].r = i+;
}
l[n-].r=;
for(int i=;i<n;i++)
q.push(l[i]);
LL ans = ;
while (!q.empty()){
A a = q.top();
q.pop();
if (mark[a.pos]) continue;
ans += l[a.pos].dis;
if (--k == ) break;
LL w = -l[a.pos].dis;
if (l[a.pos].l){
w += l[l[a.pos].l].dis;
mark[l[l[a.pos].l].pos] = ;
l[a.pos].l = l[l[a.pos].l].l;
if(l[a.pos].l)
l[l[a.pos].l].r = a.pos;
}
else
mark[a.pos] = ;
if (l[a.pos].r){
w += l[l[a.pos].r].dis;
mark[l[l[a.pos].r].pos] = ;
l[a.pos].r = l[l[a.pos].r].r;
if(l[a.pos].r)
l[l[a.pos].r].l = a.pos;
}
else
mark[a.pos] = ;
if (mark[a.pos] == ){
if (l[a.pos].r) l[l[a.pos].r].l = ;
if (l[a.pos].l) l[l[a.pos].l].r = ;
}
else{
l[a.pos].dis = w;
q.push(l[a.pos]);
}
}
printf ("%lld", ans);
return ;
}

贪心:SPOJ Backup Files的更多相关文章

  1. Python Backup Files

    近来书写 Python 脚本进行替换以前的 shell 脚本,发现 Python 优于 shell 的最直观的一点Python 结构明了,可读性高(本人认为)在此做一些记录 本次记录利用 Python ...

  2. [PowerShell] Backup Folder and Files Across Network

    ## This Script is used to backup folder/files and delete the old backup files. ## Author: Stefanie # ...

  3. TFS Express backup and restore

    When we setup source control server, we should always make a backup and restore plan for it. This ar ...

  4. ORA-19815,ORA-19809 :limit exceeded for recovery files

    数据库重新启动的时候,收到了ORA-19815的错误.从错误的提示来看,是由于闪回区的空间被填满导致无法成功启动.这种情形我们通常考虑的是清除归档日志,那就直接在OS层面rm了,真的是这样吗?客官,如 ...

  5. Mysql官方文档翻译系列-7.3.1 Establishing a Backup Policy

    原文链接 (https://dev.mysql.com/doc/refman/5.7/en/backup-policy.html) 正文 To be useful, backups must be s ...

  6. SharePoint 2013 Backup Farm Automatically With a Powershell and Windows Task Schedule

    In this post,I will show you SharePoint 2013 How to Backup Farm Automatically with a PowerShell and ...

  7. Create maintenance backup plan in SQL Server 2008 R2 using the wizard

    You will need to identify how you want your maintenance plan to be setup. In this example the mainte ...

  8. [How to] ROOT, Backup & Flash (MTKDroidTools, Spflashtool, CWM)

    这是一篇来自xda论坛的文章,写得很详细,很有用,以下是原文: Hi This is a guide to ROOT, backup and flash your MTK65xx or Other d ...

  9. [转]How to Use xp_dirtree to List All Files in a Folder

    本文转自:http://www.sqlservercentral.com/blogs/everyday-sql/2012/11/13/how-to-use-xp_dirtree-to-list-all ...

随机推荐

  1. ASP.NET Boilerplate 邮件类使用

    在系统我们自定一个 MySettingProvider,并添加到配置集合中,定义一些邮件参数覆盖默认参数,然后通过IOC容器得到SmtpEmailSender实例,调用send方法就实现了,实现代码如 ...

  2. HTTP协议 流程图

  3. CentOS安装memcached及配置php的memcache扩展

    遇到的问题: 这个问题主要是linux服务器安装memcached服务后,phpinfo信息没有memcache扩展,所以主要是给php安装memcache扩展,教程中是安装memcache扩展,我认 ...

  4. GridView中某一列值的总和(web)

    protected void GridView1_RowDataBound(object sender, GridViewRowEventArgs e)    {        if (e.Row.R ...

  5. .net开发windows服务

    最近一个月都异常的繁忙,项目进度非常的紧,回头看看自己的blog,整整一个5月都没有一篇文章,非常惭愧,现在补几篇文章,介绍一下我最近关注的技术.这篇文章将介绍Windows服务程序的开发.摘要:本文 ...

  6. iOS 去除导航栏下的黑线

    Swift: navigationController?.navigationBar.setBackgroundImage(UIImage(), forBarMetrics: .Default) na ...

  7. ID选择器

    在很多方面,ID选择器都类似于类选择符,但也有一些重要的区别: 1.为标签设置id="ID名称",而不是class="类名称". 2.ID选择符的前面是井号(# ...

  8. javascript--烟火效果

    <!DOCTYPE html> <html> <head> <meta charset="UTF-8"/> <meta nam ...

  9. photoshop中rgb与索引模式的区别

    RGB: 是色域最宽广的颜色模式它是一种光色模式不难理解  你的显示器 就是根据这个模式来显示图像的同样 我们在自然界中 看到的所有五颜六色的东西都是吸收了不同颜色的光 而得到的所以这一颜色模式 与我 ...

  10. 转载-Linux下svn搭建配置流程

    Linux下svn搭建配置流程     一.    源文件编译安装.源文件共两个,为: 1.   下载subversion源文件 subversion-1.6.1.tar.gz http://d136 ...