Problem
Mary likes playing with rubber bands. It’s her birthday today, and you have gone to the rubber band shop to buy her a gift.
There are N rubber bands available in the shop. The i-th of these bands can be stretched to have any length in the range [Ai, Bi], inclusive. Two rubber bands of range [a, b] and [c, d] can be connected to form one rubber band that can have any length in the range [a+c, b+d]. These new rubber bands can themselves be connected to other rubber bands, and so on.
You want to give Mary a rubber band that can be stretched to a length of exactly L. This can be either a single rubber band or a combination of rubber bands. You have M dollars available. What is the smallest amount you can spend? If it is impossible to accomplish your goal, output IMPOSSIBLE instead.
Limits
1 ≤ T ≤ 100.
1 ≤ Pi ≤ M.
1 ≤ L ≤ 10000.
1 ≤ Ai ≤ Bi ≤ 10000.
Small dataset
1 ≤ N ≤ 10.
1 ≤ M ≤ 100.
Large dataset
1 ≤ N ≤ 1000.
1 ≤ M ≤ 1000000000.
分析:
- 小数据应该可以通过暴力解决,对所有的N个rubber bands,枚举出所有2^N种子集,并判断每种组合是否满足长度要求,并选取其中最小花费即可。
- 但对于大数据来说 2^N太大了。选用动态规划来解决, 这边也参考了scoreboard上前几位作者的代码。
动态规划:
- 设数组dp[L+1],其中dp[i]表示要达到长度为i的rubber band所需最少花费。
- 对每个rubber band按顺序考虑到数组dp中去,即首先考虑只有前1个band时dp的状态,再考虑只有前2个band时dp的状态,……,最后考虑前N个band时的dp状态,考察dp[L]是否小于预算M即可。
- 初始时将所以dp元素设为MAX, dp[0]=0
- 当考虑第i个rubber band进入dp数组时,因为其拉伸长度可以达到[Ai,Bi], 所以对dp[j]来说,只需取dp数组中下标为j-Bi到j-Ai这一段中最小值+Pi 来决定是否更新当前dp[j]。(注意j-Bi到j-Ai不要越界)
- 当第i个rubber band进入考虑范围时,按上述更新一遍dp数组即可。
但是, 对每个j查找“dp数组中下标为j-Bi到j-Ai这一段中最小值”很耗时,考虑到其实是一个滑动窗口中的最小值,可借鉴leetcode239减少复杂度。 但一开始用deque实现比较耗时,后来直接用数组实现就变快了。
代码:Github
#include <iostream>
#include <math.h>
#include <vector>
#include <algorithm>
#include <deque>
using namespace std;
typedef long long ll;
int T;
int main() {
FILE *fin, *fout;
fin=fopen("D-large-practice.in", "r");
fout = fopen("output", "w");
fscanf(fin, "%d", &T);
for (int kk = 1; kk <= T; kk++) {
int num, money, length;
fscanf(fin, "%d %d %d", &num, &money, &length);
vector<ll> dp(length + 1, INT_MAX);
dp[0] = 0;
for (ll i = 0; i < num; i++) {
int a, b;
int p;
fscanf(fin, "%d %d %d", &a, &b, &p);
//寻找长度为b-a+1的sliding window中最小值的下标 记录到que中
vector<int> que(length + 1, 0);
int start = 0;
int end = 0;
for (int k = a; k < b&&length - k >= 0; k++) {
while (end > start && dp[que[end - 1]] >= dp[length - k]) end--;
que[end++] = length - k;
}
for (ll j = length; j >= 0; j--) {
if (j - b >= 0) {
while (end > start && dp[que[end - 1]] >= dp[j - b]) end--;
que[end++] = j - b;
}
if (start < end) dp[j] = min(dp[j], dp[que[start]] + p);
if (dp[start] >= j - a) start++;
}
}
if (dp[length] <= money)
fprintf(fout, "Case #%d: %d\n", kk, dp[length]);
else fprintf(fout, "Case #%d: IMPOSSIBLE\n", kk);
}
}