how to add 3 matrices with one dimension same for all and other dimension different?

If I have

A=[1;
   2;
   3]
B=[2 3 4;
   5 6 7;
   8 9 0]
C=[3 4 5 6;
   9 8 7 6;
   5 4 2 1]

all 3 matrices have 3 rows but columns different. I want to add elements of rows of all matrices in this way for example I want

D=[(1+2+3) (1+2+4) (1+2+5) (1+2+6) (1+3+3) (1+3+4) (1+3+5) (1+3+6) (1+4+3) (1+4+4) (1+4+5) (1+4+6);
   (2+5+9) (2+5+8) (2+5+7) (2+5+6) (2+6+9) (2+6+8) (2+6+7) (2+6+6) (2+7+9) (2+7+8) (2+7+7) (2+7+6);
   (3+8+5) (3+8+4) (3+8+2) (3+8+1) (3+9+5) (3+9+4) (3+9+2) (3+9+1) (3+0+5) (3+0+4) (3+0+2) (3+0+1)]

[John Chilleri]

Do you need this generalized to any size matrix, or will it strictly be with three matrices of these exact sizes?

[summyia qamar]

to any kind..although my specific problem is

A=64x1
B=64x1806
C=64x4602

 so I need a generalized version

Answer 1

[Niels]

[rows,col_B]=size(B);
[~,col_C]=size(C);
result=zeros(rows,col_B*col_C);

for i=1:col_B
    for  j=1:col_C
        result(:,(i-1)*col_C+j)=A+B(:,i)+C(:,j);
    end
end

[summyia qamar]

I am trying this for

A=64x1
B=64x1806
C=64x4602

but getting error Maximum variable size allowed by the program is exceeded.

[John Chilleri]

The result you're trying to produce appears to be too large, it's going to be around 64 x 4602*1806.

[summyia qamar]

yes and each value is not a single integer but its like 2345+4583+2384 etc.Can I post the complete code?

[Niels]

So you need at least 4,26GB free memory. Since you do not have an error occurs.

[summyia qamar]

means this calculation cannot be done

[John Chilleri]

Perhaps make a post with your data asking someone to run the computation for you (requiring as Niels said, 4.26 GB). I have a laptop, but I'm certain some people have much better devices.

[Niels]

in total u need more than 4,26GB, 4,26GB only for the result to be saved in the memory, but the other variables are not to mention in comparison to result

[Niels]

ah you even may have the required memory but maybe you just need to set matlabs limit higher...

Requested 64x8311212 (4.0GB) array exceeds maximum array size preference. Creation of arrays greater than this limit may take a long time and cause MATLAB to become unresponsive.

[summyia qamar]

I'm posting another question for trying to minimize the calculations in my code so that these numbers can be minimized

[Niels]

that error occurs even before you even start your calculations

[Niels]

the result will be that large, so even if the calculations are minimzed, the problem with the size of the result is still the same.

this might help you o set the limit higher https://de.mathworks.com/help/matlab/matlab_env/set-workspace-and-variable-preferences.html

Answer 2

[John Chilleri]

Hello,

I only tested this with your example and it worked, but I think it's generally correct:

%
% Strangely adding matrices
%
function D = strangelyAddingMatrices(A,B,C)
n = size(C,2);
nB = size(B,2);
for i = 1:n
  C(:,i) = C(:,i)+A;
end
D = [];
for i = 1:nB
  C2 = C;
  for j = 1:n
      C2(:,j) = C2(:,j)+B(:,i); 
  end
  D = [D C2]; 
end
end

Hope this helps!

[John Chilleri]

You would need to save it as a function then call,

D = strangelyAddingMatrices(A,B,C);

from the command window.

[summyia qamar]

ok I'm trying this. let you know about it

[John Chilleri]

Alternatively, you can copy and paste the body of the function to the command window:

n = size(C,2);
nB = size(B,2);
for i = 1:n
  C(:,i) = C(:,i)+A;
end
D = [];
for i = 1:nB
   C2 = C;
   for j = 1:n
      C2(:,j) = C2(:,j)+B(:,i); 
   end
   D = [D C2]; 
end

or you could use Neils solution below which also works!

[summyia qamar]

getting error of applying + on matrices

[John Chilleri]

That's strange, I ran it on randomly generated matrices of the same size as yours and it worked, although it slowed my computer way down as it requires so much memory (as was mentioned below).

[summyia qamar]

how can I reduce the memory consumption?

[John Chilleri]

Other than running it on a better device or asking someone to run it for you, I can only think of computing it by block.

In other words, how will you use this resultant matrix? If you are only going to use a piece of it at a time, then you can compute the necessary subblock of that matrix as needed. Or you can intelligently generate a portion at a time as required for calculations, then overwrite that portion to have eventually used the whole matrix in the calculation.

[summyia qamar]

actually these A,B,C are resultants of different calculations and I want to get minimum sum of all A,B,C thats why I want to sum up them

[John Chilleri]

What do you mean by minimum sum? Like which columns summed produce the minimum? Or something else?