poj 3017 Cut the Sequence

http://poj.org/problem?id=3017

dp+单调队列

算法的理论时间复杂度应该还是接近 O(n^2)  但为什么过得还挺快呢 应该是后台数据的问题吧

代码：

#include<iostream>
#include<cstdio>
#include<cstring>
#include<string>
#include<map>
#include<vector>
#include<stack>
#include<set>
#include<map>
#include<queue>
#include<deque>
#include<algorithm>
#include<cmath>
#define LL long long
//#pragma comment(linker, "/STACK:1024000000,1024000000")
using namespace std;
const int INF=0x3f3f3f3f;
const int N=400005;
LL a[N];
int n;
LL m,ans[N];
int dq[N],il,ir;
int main()
{
    //freopen("data.in","r",stdin);
    scanf("%d %I64d",&n,&m);
    for(int i=1;i<=n;++i)
    scanf("%I64d",&a[i]);
    for(int i=1;i<=n;++i)
    if(a[i]>m)
    {printf("-1\n");return 0;}
    il=0;ir=-1;
    LL sum=0;
    a[0]=0;
    ans[0]=0;
    int l=1;
    for(int i=1;i<=n;++i)
    {
        while(il<=ir&&a[dq[ir]]<=a[i])
        --ir;
        dq[++ir]=i;
        sum+=a[i];
        while(sum>m)
        {
            sum-=a[l];
            ++l;
        }
        while(dq[il]<l)
        {++il;}
        if(l==1)
        {ans[i]=a[dq[il]];continue;}
        ans[i]=ans[l-1]+a[dq[il]];
        for(int j=il;j<ir;++j)
        {
            int k=dq[j];
            ans[i]=min(ans[i],ans[k]+a[dq[j+1]]);
        }
    }
    printf("%I64d\n",ans[n]);
    return 0;
}