Do you mean to generate a square matrix whose elements are the sum of its subscripts? If so:
If you're using R2016b or newer, you can do
s=5; % size of matrix
A=(1:s) + (1:s)'
otherwise, you can do this in place of line 2:
A=bsxfun(@plus,(1:s),(1:s)')
Of course, this could be extended to non-square matrices:
s=[4 5]; % size of matrix
A=(1:s(2)) + (1:s(1))'
