Originally posted by: Rajasekhar Kadambur

Hi,

How to define unstructured decision variables?

A is 4X5 matrix

A11    A21    A31     A41

A21    A22    A32     A42

A31      --       A33     A43

A41      --        --         A44

  --        --         --        A45

one way of doing is to take a structured array and define A51,A23,A24,A25,A34,A35 == 0.

For large data, this process will be very painful.

could you please suggest how to implement unstructured decision variables.

Thanks in advance.

Originally posted by: davidoff

I don't know if I understood it clearly, but here is a way to flatten the matrix into a single array of variables.

 dvar float x[1..2][1..2] in 0..5;
 dvar float xflat[0..3] = all(i in 1..2,j in 1..2) x[i,j];
 
 maximize sum(i in 0..3) xflat[i];

Be carefull that the indexation of xflat must begin with 0 (or I don't know how to index in a different way)

Hope this helps

David