Index out of bounds for simple nested for loop

Hi everybody,

This function works fine when I input P and Q as matrices with equal number of rows and columns but not when they are different sizes (for example, P = [1 2 3 4; 5 6 7 8] and Q = [ 1 1 1; 2 2 2; 3 3 3; 4 4 4].

Can someone point me in the right direction as to what I can fix so that this function will work properly? Thanks.

function [M] = myMatMult (P, Q)
% function that multiplies matrices P and Q

[nRow, nCol] = size(P);
[nRow_1, nCol_1] = size (Q);

if nCol ~= nRow_1
    disp 'error';
end

M = zeros(nRow, nCol_1);

for i = 1:nRow
    for j = 1:nCol_1
        for k = 1: nCol 

             M(i,j) =   M(i, j) + P(i,k)* Q(j,k);

          end

      end

end

end

Answer 1

[Geoff Hayes]

Allie - since you are performing your own version of matrix multiplication, if you are multiplying P*Q, then the number of columns in P must be identical to the number of rows in Q.

Now consider your multiplication step

 M(i,j) =   M(i, j) + P(i,k)* Q(j,k);

where i is from 1 to the number of rows in P, j is from 1 to the number of columns in Q, and k is from one to the number of columns in P. Both i and k are fine for indexing into P but look at the indexing into Q: we use j to access a row, and k to access a column when it should be the other way around! Since the number of columns in Q is equal to the number of rows in P and k is from one to the number of columns in P, we must use k to access a row of Q. So just change this line of code to

 M(i,j) =   M(i, j) + P(i,k)* Q(k,j);

and you should be good to continue.

[Allie]

Thank you Geoff! That did the trick.

[Geoff Hayes]

Glad it worked out!