http://acm.hdu.edu.cn/showproblem.php?pid=6058
Give you an array A[1..n]of length n.
Let f(l,r,k) be the k-th largest element of A[l..r].
Specially , f(l,r,k)=0 if r−l+1< k.
Give you k , you need to calculate ∑nl=1∑nr=lf(l,r,k)
There are T test cases.
1≤T≤10
k≤min(n,80)
A[1..n] is a permutation of [1..n]
∑n≤5∗105
Input
There is only one integer T on first line.
For each test case,there are only two integers n,k on first line,and the second line consists of n integers which means the array A[1..n]
Output
For each test case,output an integer, which means the answer.
Sample Input
1
5 2
1 2 3 4 5
Sample Output
30
题解
这是一道技巧性非常强的一道题
如果能想到从小枚举第k大的值的话这个题也就能AC了,因为比当前值小的数字对当前值的第k大无影响,我们要知道的是比当前值大的所在位置即可,利用链表保存数据,每次枚举完了对当前数字进行删除 o(1) ,那么链表里存的永远是比当前值大的数,那么只要在当前数值的位置往链表两边枚举k个值得到位置即可,有了位置不难得到答案,复杂度 o(n⋅k)
#pragma GCC optimize ("O2")
#include<stdio.h>
#include<vector>
#include<algorithm>
using namespace std;
typedef long long LL;
const int MAXS = 20 * 1024 * 1024;
char buf[MAXS], *ch;
inline void read(int &x) {
while(*ch <= 32) ++ch;
for(x = 0; *ch >= 48; ++ch) x = x * 10 + *ch - 48;
}
const int MAXN = 5e5 + 5;
int pos[MAXN], par[MAXN], nxt[MAXN], n, k;
int lst[85], rst[85], lsz, rsz;
int main()
{
fread(buf, MAXS, 1, stdin);
ch = buf;
int T, x;
read(T);
while(T--) {
read(n);
read(k);
for(int i = 1; i <= n; ++i) {
read(x);
pos[x] = i;
par[i] = i - 1;
nxt[i] = i + 1;
}
LL ans = 0;
for(int i = 1; i <= n; ++i) {
lsz = rsz = 0;
int it = pos[i];
while(0 < it && lsz < k) {
lst[lsz++] = it - par[it];
it = par[it];
}
it = pos[i];
while(it <= n && rsz < k) {
rst[rsz++] = nxt[it] - it;
it = nxt[it];
}
int lp = lsz - 1, rp = k - lsz;
while(rp < rsz && i + rp <= n) {
ans += 1LL * i * lst[lp--] * rst[rp++];
}
par[nxt[pos[i]]] = par[pos[i]];
nxt[par[pos[i]]] = nxt[pos[i]];
}
printf("%lld\n", ans);
}
return 0;
}