poj3181 dp 优化递推关系+高精度

 

如题：http://poj.org/problem?id=3181

Dollar Dayz

Time Limit: 1000MS		Memory Limit: 65536K
Total Submissions: 4850		Accepted: 1846

Description

Farmer John goes to Dollar Days at The Cow Store and discovers an unlimited number of tools on sale. During his first visit, the tools are selling variously for $1, $2, and $3. Farmer John has exactly $5 to spend. He can buy 5 tools at $1 each or 1 tool at $3 and an additional 1 tool at $2. Of course, there are other combinations for a total of 5 different ways FJ can spend all his money on tools. Here they are:

        1 @ US$3 + 1 @ US$2

        1 @ US$3 + 2 @ US$1

        1 @ US$2 + 3 @ US$1

        2 @ US$2 + 1 @ US$1

        5 @ US$1

Write a program than will compute the number of ways FJ can spend N dollars (1 <= N <= 1000) at The Cow Store for tools on sale with a cost of $1..$K (1 <= K <= 100).

Input

A single line with two space-separated integers: N and K.

Output

A single line with a single integer that is the number of unique ways FJ can spend his money.

Sample Input

5 3

Sample Output

5

Source

USACO 2006 January Silver

 

 

 

思路：dp[i][j]表示i任意划分，最大的数也不超过j的划分数。

dp[i][j]=∑(k=0,i/j) dp[i-k*j][j-1]   相当于从i中取出任意个j，剩下的最大的数不超过j-1。

注意到dp[i][j]由两种状态推出，一种是最大值小于j，一种是最大值等于j。

因此优化递推关系得：

dp[i][j]=dp[i][j-1]+dp[i-j][j]。

 

数据量很大，连long long 都会超，因此使用高精度。

dp[i][j]:保存的是当前(i,j)状态的进位。

t[i][j]:保存的是取模的余数。（64位int为19位）

 

#include<iostream>
#include<cstdio>
#include<cstring>
using namespace std;
long long dp[1005][105],t[1005][105],MOD;

int main()
{
 int n,m;
 cin>>n>>m;
 int i,j;
 for(i=0;i<=m;i++)
 {
  t[0][i]=1;
 }
 MOD=1;
 for(i=0;i<18;i++)
  MOD=MOD*10;
 for(i=1;i<=n;i++)
 for(j=1;j<=m;j++)
   if(i>=j)
   {
    dp[i][j]=dp[i-j][j]+dp[i][j-1]+(t[i-j][j]+t[i][j-1])/MOD;
    t[i][j]=(t[i-j][j]+t[i][j-1])%MOD;
   }
   else
   {
    dp[i][j]=dp[i][j-1];
    t[i][j]=t[i][j-1];
   }
 if(dp[n][m])
  printf("%I64d",dp[n][m]);
 printf("%I64d\n",t[n][m]);
 return 0;
}