CAT arguments dimensions are not consistent.

Why am I getting this error:

>> fpi1
??? Error using ==> horzcat
CAT arguments dimensions are not consistent.

Error in ==> fpi1 at 17
[x' g(x)' e' rat']

I want to output all of the 4 values and x and g(x) would help me to see the roots.

%Program 1.2 Fixed-Point Iteration
%Computes approximate solution of g(x)=x
%Input: inline function g, starting guess x0, 
%       number of steps k
%Output: Approximate solution xc
g=@(x) (5-3*x.^3)/(11.*x+11);
x(1)=1;steps=20;
for i=1:steps
  x(i+1)=g(x(i));
end
r=x(steps+1);
e=x-r;
for i=1:steps
  rat(i+1)=e(i+1)/e(i);
end
rat(1)=0;
[x' g(x)' e' rat']

Answer 1

[Star Strider]

You need to do element-wise division as well. See if this change solves the problem:

g=@(x) (5-3*x.^3)./(11.*x+11);

Answer 2

[Kirby Fears]

Your last statement is attempting to combine arrays x, g(x), e, and rat into a single array. To combine the arrays correctly, you have to do so over a dimension that is the same size in each array.

If you remove the transpose operators, the last line becomes

 [x g(x) e rat]

and this works fine. All of these arrays have a size of 1 in dimension 1 (e.g. they have 1 row each).

I'm not sure this is the operation you were intending to perform. If you're trying to store all of these values into a single structure to pass somewhere else, you can combine them into a struct.

 output.x=x;
 output.gx=g(x);
 output.e=e;
 output.rat=rat;

On a side note, you should not use "e" as a variable name because it's generally associated with the mathematical constant e. Also, 1e10 in Matlab means 1*10^10. You don't want any ambiguity in your code.