New to programming can someone tell me where I am going wrong?

[workmanj]

Hi,
I am trying to write a program to handle some simple numerical integration schemes and I keep having certain errors. Whenever I try to convert the integer types from float to double the program stops working, I am getting accuracy errors using float so I would like to know what I am doing wrong. Also on the monte carlo routine I am trying to generate a random number between a and b and then convert it back to a floating point but if you run the code you can tell it doesnt work. Any general advice in addition would be helpful. I am trying to self teach myself c for grad school and have no formal background in any languages.

Please see code below:

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



/* use of variable declaration outside of main causes them to be global*/

void simp(void);
void trap(void);
void monte(void);
float f1(float x);
float f2(float x);

main ()
{
int choice;

printf(&quot;\nHello there and welcome to Jared's integration routine\n&quot;

while(choice != 4)
{
printf(&quot;\nWould you prefer to use 1-The Trapezoidal Method (x^3) 2-Simpson's Method (e^(-x)) 3-Monte Carlo 4-exit\n&quot;
scanf(&quot;%d&quot;, &choice);
switch(choice)
{
case 1: /*Trapezoidal Method for x^3*/
trap();
break;

case 2: /*Simpson's Method*/
simp();
break;


case 3: /*Monte Carlo Method*/
monte();
break;

}
}
printf(&quot;\n Goodbye &quot;
}


/*Trapezoidal method*/

void trap(void)
{
float x1, x2, sumt, a, b, step;

printf(&quot;\nWhat is your starting point?\n&quot;
scanf(&quot;%f&quot;, &a);

printf(&quot;\nWhat is your end point?\n&quot;
scanf(&quot;%f&quot;, &b);

printf(&quot;\nWhat is your step size?\n&quot;
scanf(&quot;%f&quot;, &step);

sumt = 0.0;
x1 = a;

printf(&quot;%f %f %f&quot;, a, b, step);

for (x2 = (a + step) ; x2 <= b; x2 = (x2 + step))
{
sumt += ((step)*(f1(x1) + f1(x2)))/2.0;
x1 = x2;
}
printf(&quot;\n The answer is %f&quot;, sumt);
return;
}



/*Simpson's Rule*/

void simp(void)

{
float sums, x1, x2, a, b , step;
int i;

printf(&quot;\nWhat is your starting point?\n&quot;
scanf(&quot;%f&quot;, &a);

printf(&quot;\nWhat is your end point?\n&quot;
scanf(&quot;%f&quot;, &b);

printf(&quot;\nWhat is your step size?\n&quot;
scanf(&quot;%f&quot;, &step);

sums = (f2(a) + f2(b));
x1 = a;

for (x2 = (x1 + step); x2 <= b; x2 = (x2 + step))
{
sums += (2.0*f2(x2));
}

sums = (sums * step)/2.0;
printf(&quot;\n The answer is %f&quot;, sums);
return;

}

/*Monte Carlo Method*/

void monte(void)
{
float summ, yr, f, c = 0.0, rx, rxn, ry, ryn;
int rand();
int i, n, a, b, rxi;

printf(&quot;\nWhat is your starting point?\n&quot;
scanf(&quot;%f&quot;, &a);

printf(&quot;\nWhat is your end point?\n&quot;
scanf(&quot;%f&quot;, &b);


printf(&quot;\nHow many iterations would you like?\n&quot;
scanf(&quot;%d&quot;, &n);

for (i=1; i<=n; ++i)
{

rxi = rand()%(b-a+1)+a;
printf(&quot;\n%d&quot;, rxi);
rxn = rxi/1.0;
printf(&quot;\n%f&quot;, rxn);
yr = f1(rxn);
printf(&quot;\n%&quot;, yr);
ry = rand()/32768.0;
ryn = (yr * ry);
c += (ry <= yr) ? 1.0 :0.;
}
f = (float) c/i;
summ = f*f1(b - a);
printf(&quot;\n The answer is %f&quot;, summ);




return;
}



/* function to evalaute by trapezoidal method*/
float f1(float x)
{
return ( x * x * x );
}

/* function to evalaute by simpson's method*/
float f2(float x)
{
return (exp(-x));
}

 

[Salem]

Well there are a few errors which need to be fixed

$ gcc -Wall -W hello.c -lm
hello.c:17: warning: return type defaults to `int'
hello.c: In function `main':
hello.c:44: warning: control reaches end of non-void function
hello.c: In function `simp':
hello.c:84: warning: unused variable `i'
hello.c: In function `monte':
hello.c:118: warning: float format, different type arg (arg 2)
hello.c:121: warning: float format, different type arg (arg 2)
hello.c:135: warning: spurious trailing `%' in format
hello.c:135: warning: too many arguments for format
hello.c:113: warning: unused variable `rx'

17 - say int main ( )
34 - say return 0;
84 - just delete the unused var
118 - make the format consistent with the variable - %d for int say
121 - ditto
135 - &quot;\n%&quot; is not a format - perhaps &quot;%d\n&quot; ?



Also in monte(), you want to delete this line
int rand();
There is no need to re-prototype functions which are already prototyped by those include files.

 

[workmanj]

Thanks,
I realized I was scaning doubles using %f and not %lf and this fixed that problem.

Would you by any chance know how to write a random number generator which creates double or float values and computes numbers in a prespecified interval?

Thanks,
Jared

 

[Salem]

Well

double r = (double)rand() / RAND_MAX;

produces pseudo random numbers in the range [0..1.0]
After that, you can scale it to whatever you want.

http://directory.google.com/Top/Computers/Algorithms/Pseudorandom_Numbers/?tc=1

You should be aware that the standard library rand() is pretty much useless for serious statistical purposes.

This one seems like a good one for statistical purposes (from reading the words - make your own mind up)

http://www.math.keio.ac.jp/~matumoto/emt.html

It is available as a download