Codeforces Round #317 [AimFund Thanks-Round] (Div. 2) D. Minimization（dp）

题意:

$给定一个N\le3\times10^5的序列, 给定一个1\le K\le min(5000,n-1)$
$通过重排原序列, 求最小的\sum_{i=1}^{n-k}|A_i-A_{i+k}|$

分析:

$观察这个式子, 里面其实就是|A_j-A_i|, j\equiv i\ mod\ k$
$那么我们把原序列按照模k的余数分类,总共有k个子序列, 0\sim k-1$
$发现其中有n\%k个子序列的长度为n/k+1,n-n\%k个长度为n/k$
$假设其实一个序列是b_1,b_2,...,b_m这样排列值最优的$
$它的贡献就是E_i=|b_2-b_1|+|b_3-b_2|+|b_4-b_3|+\cdots+|b_m-b_{m-1}|$
$让这样的贡献最小, 显然我们应该让这个序列的相邻数的差距最近$
$那么这样的序列一定是原序列有序排序后, 其中连续的一部分$
$这样之后b_1\le b_2\le \cdots \le b_m$
$显然E_i=b_2-b_1+b_3-b_2+b_4-b_3+\cdots+b_m-b_{m-1}=b_m-b_1$
$那么原问题就转化为有序序列中选取n\%k个长连续子序列, n-n\%k个短连续子序列的最优解$
$dp[large][small]:=选取large个长连续子序列, small个短连续子序列的最优解$
$转移就枚举选取的这个是长还是短就可以了$
$时间复杂度为O(nlogn+k^2)$
$注意细节。$

代码:

//
//  Created by TaoSama on 2016-02-04
//  Copyright (c) 2016 TaoSama. All rights reserved.
//
#pragma comment(linker, "/STACK:1024000000,1024000000")
#include <algorithm>
#include <cctype>
#include <cmath>
#include <cstdio>
#include <cstdlib>
#include <cstring>
#include <iomanip>
#include <iostream>
#include <map>
#include <queue>
#include <string>
#include <set>
#include <vector>

using namespace std;
#define pr(x) cout << #x << " = " << x << "  "
#define prln(x) cout << #x << " = " << x << endl
const int N = 3e5 + 10, INF = 0x3f3f3f3f, MOD = 1e9 + 7;

typedef long long LL;
int n, k, a[N];
LL f[5005][5005];

int main() {
#ifdef LOCAL
    freopen("C:\\Users\\TaoSama\\Desktop\\in.txt", "r", stdin);
//  freopen("C:\\Users\\TaoSama\\Desktop\\out.txt","w",stdout);
#endif
    ios_base::sync_with_stdio(0);

    while(scanf("%d%d", &n, &k) == 2) {
        for(int i = 1; i <= n; ++i) scanf("%d", a + i);
        sort(a + 1, a + 1 + n);
        int len = n / k, large = n % k, small = k - large;
        memset(f, 0x3f, sizeof f);
        f[0][0] = 0;
        for(int i = 0; i <= large; ++i) {
            for(int j = 0; j <= small; ++j) {
                int st = i * (len + 1) + j * len + 1;
                if(i < large) {
                    int largeCost = a[st + len] - a[st];
                    f[i + 1][j] = min(f[i + 1][j], f[i][j] + largeCost);
                }
                if(j < small) {
                    int smallCost = a[st + len - 1] - a[st];
                    f[i][j + 1] = min(f[i][j + 1], f[i][j] + smallCost);
                }
            }
        }
        printf("%I64d\n", f[large][small]);
    }
    return 0;
}