Working on a function that performs gaussian elemination

I am working on a matlab function that can perform gaussian elimination on a matrix of any size

Currently I have this

function x = gauss(A,b)
% This function performs the Gauss elimination without pivoting
% 
% x = GAUSS(A, b)

[n,n] = size(A); 

% Check for zero diagonal elements
if any(diag(A)==0)
    error('Division by zero will occur; pivoting not supported')
end

% Forward elimination
for row=1:n-1
    for i=row+1:n
        factor = A(i,row) / A(row,row);
        for j = row:n
            A(i,j) = A(i,j) - factor*A(row,j);
        end
        b(i) = b(i) - factor*b(row);
    end
  A_and_b =  [A b]
end

% Backward substitution
x(n) = b(n) / A(n,n);
for row = n-1:-1:1
    sums = b(row);
    for j = row+1: n
        sums = sums - A(row,j) * x(j);
    end
    x(row) = sums / A(row,row);
end

I have my function a and b defined in the command window but when I run the function in the command window I get an error message saying

Error using size
Too many output arguments.

Error in gauss (line 5)
[n,n] = size(a,b)

I am not sure how to fix this as I am not very familiar with matlab. How do I get rid of this error?

[John Smith]

Thank you very much, I am now able to get an output but my matrix looks like this

a_and_b =

       1     2    -1     0     1
       0     0     0    -1    -3
       0   NaN   NaN   Inf   Inf
       0   NaN   NaN   NaN   NaN

ans =

     NaN   NaN   NaN   NaN

Any suggestions on how to correct this and turn the output into real values?

Answer 1

[Walter Roberson]

Your error message shows

[n,n] = size(a,b)

where you have passed two arguments in to size(). When you pass two arguments in to size() you can only have one output.

The code you posted does it right,

[n, n] = size(A)

[John Smith]

Thank you very much, I am now able to get an output but my matrix looks like this

a_and_b =

       1     2    -1     0     1
       0     0     0    -1    -3
       0   NaN   NaN   Inf   Inf
       0   NaN   NaN   NaN   NaN

ans =

     NaN   NaN   NaN   NaN

Any suggestions on how to correct this and turn the output into real values?

[Steven Lord]

What were the A and b matrices with which you called your function?

[John D'Errico]

To answer your followup question, why you get this:

a_and_b =

       1     2    -1     0     1
       0     0     0    -1    -3
       0   NaN   NaN   Inf   Inf
       0   NaN   NaN   NaN   NaN

It looks like your "A" matrix had a zero pivot. The NaNs result from a 0/0 operation. (A guess, since we do not know the matrix A.) But that is what I would expect to see if you got that result from a Gaussian elimination that did not employ pivoting.

This may be an indication that your matrix is singular, or it merely may be a carefully chosen matrix that is non-singular. (I am fairly sure it was carefully chosen by your instructor.)

[John Smith]

[John Smith]

That is the matrix I am working with, do you think I need to implement code for pivoting to make it work because I know the result is intended to produce values. Thanks again