求助一个关于Subteck2的问题

为什么要把 f_i 初始化为 0xcf 呢？

Subteck2 全错代码：

#include <bits/stdc++.h>
using namespace std;
int n, l, r, ans = -10000000;
int a[200001], f[200001];
int q[200001]; //优先队列 
int main()
{
    cin >> n >> l >> r;
    for(int i = 0; i <= n; i++)
        cin >> a[i];
    //memset(f, 0xcf, sizeof(f));
    //f[0] = 0;
    int head = 0, tail = -1;
    for(int i = l; i <= n; i++)
    {
        while(head <= tail && q[head] < i - r) head++;
        while(head <= tail && f[i - l] > f[q[tail]]) tail--;
        q[++tail] = i - l;
        f[i] = f[q[head]] + a[i];
    }
    for(int i = n + 1 - r; i <= n; i++)
        ans = max(ans, f[i]);
    cout << ans;
    return 0;
}

Subteck2 全对代码：

#include <bits/stdc++.h>
using namespace std;
int n, l, r, ans = -10000000;
int a[200001], f[200001];
int q[200001]; //优先队列 
int main()
{
    cin >> n >> l >> r;
    for(int i = 0; i <= n; i++)
        cin >> a[i];
    memset(f, 0xcf, sizeof(f));
    f[0] = 0;
    int head = 0, tail = -1;
    for(int i = l; i <= n; i++)
    {
        while(head <= tail && q[head] < i - r) head++;
        while(head <= tail && f[i - l] > f[q[tail]]) tail--;
        q[++tail] = i - l;
        f[i] = f[q[head]] + a[i];
    }
    for(int i = n + 1 - r; i <= n; i++)
        ans = max(ans, f[i]);
    cout << ans;
    return 0;
}