Calculation error

[Vigii]

Hi,

I'm trying to calculate values of array b and d.

I have checked that all the values which are assigned to arrays a and b are positive values.

But when I try to calculate using this loop the values of o is alternatively changing from positive to negative.

there is some kind of error within this part that I'm unaware.

Please guys help me to fix this error.

   for(int i=3;i<j;i++)
  {  
          o=a[i]/b[i-1];
          b[i]=b[i]-(o*c[i-1]);
          d[i]=d[i]-(o*d[i-1]);
  }

[Chervil]

You show the code, which is your proposed solution. But you didn't state what the original problem was, so it's not possible to know whether this is correct or not.

[Vigii]

Hi Chervil,

I have posted you entire code.

I tried to fix this bug. it works well till line 52 .

The problem is with array b[i-1]. when i try to calculate and print b[i-1].

It prints alternative negative numbers. But I didn't assign any negative values to any of arrays.

I don't understand how does it prints a negative value for b[i-1].

But if I'm taking b[i] instead of b[i-1]. It shows perfect value which has to be same for b[i-1] also.

please kindly go through the code. Let me know where is the error.

Thanks in Advance

#include <iostream>
#include <conio.h>
#include <algorithm>
#include <fstream>
#include <iomanip>
#include <sstream>

using namespace std;

int main()
{
    double dt;
    int j=41,n=45,k;
    double uy[45],y[45],a[n],b[n],c[n],d[n];
    double dy=0.001,mu=0.000217,f;
    float o;
    cout<<"Enter time step interval of your choice : ";
    cin>>dt;
    
         for(int i=1;i<=j;i++)
            if (i==1)
                {   
                    y[i]=40;
                }
                else
                { y[i]=0;
                }

f=mu*dt/(dy*dy);
  
  for(int i=2;i<j;i++)
  {
          if (i==2)
          {
             a[i]=f;
             b[i]=(-2*f)+1;
             c[i]=f;
             d[i]=-1-(a[i]*y[i-1]); 
          }else if (i==j-1)
          {
             a[i]=f;
             b[i]=(-2*f)+1;
             c[i]=f;
             d[i]=-1-(a[i]*y[i+1]);    
          }else
          {     
             a[i]=f;
             b[i]=(-2*f)+1;
             c[i]=f;
             d[i]=-1;
          } 
  }
         
  for(int i=3;i<j;i++)
  {  
          o=a[i]/b[i-1];
          b[i]=b[i]-(o*c[i-1]);
          d[i]=d[i]-(o*d[i-1]);         
  }
                      
y[j-1]=d[j-1]/b[j-1];


for (int i=j-2;i>1;i--)
{
  y[i]=(d[i]-(c[i]*y[i+1]))/b[i];

}
     
cout<<"CALCULATION COMPLETE !!!!";        
getch();
}

[TheIdeasMan]

@Vigii,

I noticed these things:

On line 29 dt is not initialised, producing garbage values for f, and for all the values in the a array.

On lines 36, 42, 48 b[i] is always negative by assignment, but could be positive given the garbage value for f.

Btw, I think you should always use doubles, not floats - mixing the two can cause problems, and there is no need to use floats (presumably not processing billions of them).

Hope all goes well.

[Vigii]

Hi,

You can see dt is read as user input on line 18.

and also the value is non-negative yielding 0.132 as a value for b[i].

Please kindly check it.

[TheIdeasMan]

Sorry, my bad.

I still think you should use doubles exclusively.

The only other suggestion I have is to use the debugger to deduce where things go wrong. using the debugger can save one heaps of time tracking down run time errors.

Hope all goes well.