POJ 3331: Factorial Calculation (High Precision Problem)

1. Overview:
   Give two numbers n, m, M is 0 to 9, ask n! How many m in the result;
   Input: (1) t: number of test cases; (2) n, m, total T lines;
   Output: n! Contains the number of M, a total of T lines.

2. Solution：
   There are several procedures:

3. Recursion is a common method to find factorial. Design a procedure, fac (n) = n × fac (n-1);
   The possible problem is that the result is too big for the computer to handle, for example, a 1000 bit number;

4. Therefore, we must use high-precision algorithm to find the factorial of a number n by processing them in arrays;

5. Judge the number of a number in each digit of a number;

6. Multiplication by High Precision Calculation
   Simulate the process of vertical multiplication of a × B, a [0] is a bit, and the result of multiplication is put into sum. Vertical multiplication is: the number of a multiplied by the number of B, the number of a multiplied by the number of B I’ll make it up again.
   Note that a [i] × B [J] corresponds to the I + J bit of the product, that is sum [i + J], which means: in the vertical formula, the value of a bit multiplied by B bit (a [0] × B [0]) is in the s [0 + 0] of the result, the value of a bit multiplied by B bit (a [0] × B [1]) is in the s [0 + 1] of the result, and so on
   Because the value of the array is not limited, we should add up all the numbers that should be in a certain bit of the result (such as I + J bits), and then process the carry uniformly. Therefore, it is divided into two steps, first multiplying by bits and then accumulating, and then modulo 10 carry each bit of the result;

for(i=0;i<1001;i++)
{
for(j=0;j<1001;j++)
{
s[i+j]=s[i+j]+a[i]*b[j];
}
}
for(i=0;i<2000;i++)
{
s[i+1]=s[i+1]+s[i]/10;
s[i]=s[i]%10;
}

3. Factorial by High Precision Calculation
   The simulation process is: for I, suppose that the result of (i-1)! Has been stored by bit in array a, then I! That is, I is multiplied by every bit in the a array starting from one bit. That is to say, starting from a [0], every bit of a in turn × I, after adding the carry (t records the carry, the initial value is 0), the remainder divided by 10 remains in this bit, and the quotient divided by 10 is used as the carry t to participate in the next calculation. The following is the core code for the problem:

int s=fac[j]*i+c;
fac[j]=s%10;
c=s/10;

5. Code:

#include
#include
using namespace std;
int fac[10005];
int main()
{
int t,n,m,c,sum,i,j,k;
scanf("%d",&t);
while(t!=0)
{
scanf("%d%d",&n,&m);
t–;
for(i=0;i<1005;i++)
{
fac[i]=0;
}
fac[0]=1;
sum=0;
for(i=2;i<=n;i++)
{
c=0;
for(j=0;j<1000;j++)
{
int s=fac[j]*i+c;
fac[j]=s%10;
c=s/10;
//cout<<s<<" “;
//cout<<fac[j]<<” “;
//cout<<c<<” “;
}
//cout<<endl<<endl;
}
//cout<<endl;
for(k=1000-1;k>0;–k)
{
if(fac[k]) // Is 0, no number
break;
}
for(i=0;i<=k;i++) // From 0 to the bits with number, judge how many M are there
{
if(fac[i]==m)
sum++;
}
printf(”%d\n",sum);
}
system(“pause”);
return 0;
}

6. Original Problem: http://poj.org/problem?id=3331