题意: 不概括了..太长了..
额第一次做这种问题 算是概率dp吗?
保存前缀项中第一个和最后一个的概率 然后每添加新的一项 就解除前缀和第一项和最后一项的关系 并添加新的一项和保存的两项的关系
这里关系指的是两者相邻会产生的额外收入(其中一个满足条件就能得到 因此公式是 2000 * (rate[a] * rate[b] + rate[a] * ( 1 - rate[b]) + rate[b] * (1 - rate[a]))
至于一开始为什么老是调不过去呢..我发现添加第三项的时候前两项是不用解除关系的...
啊啊啊啊这题我敲了大半个小时敲得都烦死了= = 不过交上去直接Pretest Pass了
代码稍显冗长.. 凑和着看吧..
#include <cstdio> #include <cmath> #include <cstring> #include <queue> #include <vector> #include <algorithm> #define INF 0x3f3f3f3f #define mem(str,x) memset(str,(x),sizeof(str)) #define STOP puts("Pause"); using namespace std; typedef long long LL; int main() { int n, p; int l, r; double rate[3], ans = 0; scanf("%d%d", &n, &p); scanf("%d%d", &l, &r); if(l % p) rate[0] = (double)(r / p - l / p) / (r - l + 1); else rate[0] = (double)(r / p - l / p + 1) / (r - l + 1); //printf("i = %d rate = %f ", 1, rate[0]); for(int i = 2; i <= n; i++){ scanf("%d%d", &l, &r); if(l % p) rate[2] = (double)(r / p - l / p) / (r - l + 1); else rate[2] = (double)(r / p - l / p + 1) / (r - l + 1); if(i == 2){ ans += (rate[0] * rate[2] + rate[0] * (1 - rate[2]) + rate[2] * (1 - rate[0])) * 2000; rate[1] = rate[2]; } else if (i == 3){ ans += (rate[0] * rate[2] + rate[0] * (1 - rate[2]) + rate[2] * (1 - rate[0])) * 2000 + (rate[1] * rate[2] + rate[1] * (1 - rate[2]) + rate[2] * (1 - rate[1])) * 2000; rate[1] = rate[2]; } else{ ans += (rate[0] * rate[2] + rate[0] * (1 - rate[2]) + rate[2] * (1 - rate[0])) * 2000 + (rate[1] * rate[2] + rate[1] * (1 - rate[2]) + rate[2] * (1 - rate[1])) * 2000 - (rate[0] * rate[1] + rate[0] * (1 - rate[1]) + rate[1] * (1 - rate[0])) * 2000; rate[1] = rate[2]; } //printf("i = %d rate = %f ans = %f ", i, rate[2], ans); } printf("%f ", ans); return 0; }