funtion don't run

F=[];
A=[];
m=5;
sum=0;
for n=100:100:2000
  for f=1:1:n
    for i=0:1:f
      p=exp(-(n-m-1)/f);
      sum=sum+(factorial(f)/(factorial(i)*factorial(f-i)))*p^i*(1-p)^(f-i)*(1-i/f)^(m+1);
      if sum<0,05;
        F=[F i];
      end       
    end
    Fm=min(F)
  end
  A=[A Fm];
end

[dpb]

>> factorial(170)
ans =
7.2574e+306
>> factorial(171)
ans =
 Inf
>>

In short, "no can do" to evaluate limits via floating point artithmetic; you'll have to approach this in a different manner entirely.

What's the actual problem to be solved?