BZOJ1076[SCOI2008] Rewards off __ probability and expectation

Description

You are playing your favorite video game and have just entered a reward pass. In this reward Sekiri, the system will randomly throw k-Times treasures,
Every time you choose to eat or not eat (you must make a choice before throwing the next treasure, and now decide not to eat the treasures can not eat later).
There are a total of n kinds of treasures, the system every time the probability of throwing these n treasures are the same and independent of each other. In other words, even the former k-1 system throws a treasure 1 (
This situation is likely to occur, although the probability is very small, the K-time to throw the probability of each treasure is still 1/n. Get the first I kind of treasures will get pi
Points, but not every treasure is free to get. The first kind treasures have a prerequisite treasure set SI. Only when all the treasures in Si have been eaten at least
Once, can eat the first kind of treasures (if the system throws a currently cannot eat treasures, equal to lose a chance in vain). Note that PI can be
To be negative, but if it is the premise of many highly divided treasures, the loss of short-term benefits to eat this negative treasure will gain greater long-term benefits. Suppose you
Take the optimal strategy and how much you can score on the average in the award. Input

The first behavior is two positive integers k and n, that is, the number and type of treasures. The following n lines describe a treasure, in which the first integer represents the score, with
The integers in turn represent the artifacts (each treasure numbered 1 to N), ending with 0. Output

Outputs a real number that retains six decimal digits, i.e. the score of the average situation under the optimal strategy. Sample Input 1 2
1 0
2 0 Sample Output 1.500000 HINT

"Scale of data"

1<=k<=100,1<=n<=15, the score is an integer in [ -10^6,10^6].

DP F[i][j] For the first time the desired transfer of the current state of J is F[i][j]+=max (f[i+1][j| ( 1<< (L-1))]+v[l],f[i+1][j])/n;

That is, select or not to transfer in turn can make F[I][J] This state will only be counted 1 times and will appear repeat if the optimal DP structure is not problematic but in the recursive or expected such models will have problems the final answer is F[1][0]

#include <iostream>
#include <cstdio>
#include <cstring>
using namespace std;
const INT maxn=1<<17;
Double F[110][MAXN];
int s[20];
int v[20];
int main () {
    //freopen ("a.in", "R", stdin);
    Freopen ("A.out", "w", stdout);
    int n,k;
    scanf ("%d%d", &k,&n);
    int x;
    for (int i=1;i<=n;i++) {
        scanf ("%d", &v[i]);
        scanf ("%d", &x);
        while (x!=0) {
            s[i]|= (1<< (x-1));
            scanf ("%d", &x);
        }
    int top= (1<<n)-1;
    for (int i=k;i>=1;i--) {for
        (int j=0;j<=top;j++) {for
            (int l=1;l<=n;l++) {
                if (s[l]&j) ==s [l])
                    F[i][j]+=max (f[i+1][j| ( 1<< (L-1))]+v[l])/double (n), f[i+1][j]/double (n));
                else f[i][j]+=f[i+1][j]/double (n);
    }} printf ("%.6lf\n", F[1][0]);
return 0;
}