c math calculation

alaios · 05-30-2004, 09:33 AM

Hi i want to calculate the following function

T=-(1/l)*ln(1-y)...

I must do this without using an ln function...but only using the log...
Any good suggestion?
Thx

eric.r.turner · 05-30-2004, 10:37 AM

You know, of course, that ln(a) == log_e(a). Remember that one of the properties of logarithms is log_b(a) == log_n(a)/log_n(b). So, ln(a) = log(a)/log(e). Here's a little C sample to show that the two are equivalent:

Code:

#include <math.h>
#include <stdio.h>

int main( int argc , char* argv[] ) {

   double a = 123.45;
   double e = exp(1.0);
   double maxDelta = 0.000000000000001;
   double delta = 1.0;

   delta = fabs( log(a) - log10(a)/log10(e) );

   if( delta < maxDelta ) {
      printf( "log_b(a) == log_n(a)/log_n(b)\n" );
   }
   else {
      printf( "log_b(a) != log_n(a)/log_n(b)\n" );
   }

   return( 0 );
}

When you compile you have to link in the math library:

Code:

gcc -o foo foo.c -lm

alaios · 05-31-2004, 06:06 AM

Do u know if the ln function exists?
I 've tried the ln(y) but it didn't work...

Now as far as your program is concerned i would say

that it doesn't seem to be working

T = (float) rand()/RAND_MAX;
//ln(a) = log(a)/log(e)
printf("*************T = %f \n",T);
temp=log(1-T) - log10(1-T)/log10(e);
T=-(1/l)*temp;

I want to calculate the T=1(1/l)*ln(1-y)

Are u sure that the e must be declared as double... I think that float is much more proper....

Thx a lot

eric.r.turner · 06-01-2004, 01:46 AM

In math.h, the natural log function (ln) is really log(). Read the man page for log().

You want double for e. Double has greater precision than float, and the exp() function returns a double. Read the man page for exp().

My little program only shows you how to calculate ln() without using the math.h log() function. Instead you use log10(), which is log base 10. The rest is up to you.

Before you use a function you are not familiar with, you really must read the documentation for that function and make sure you are using the correct argument types and return types!