Originally posted by: CPlex_NooBie

Hey! The following code snippet explains the curiosity:

{int} M=...;
tuple t
{
  int beta;
  int alp;
  int r;
};

t allData[M]=...;       
dvar int y[M][1..r][N] in 0..1;

The last line is the problem where the size of the second index depends on 'r' and the value of 'r' depends on M. Hence the solver gives an error 'circular dependency'. How do I initialize the decision variable y?

Hi,

you may either use a bigger array like:

{int} M= {1,2};

range N=1..2;
tuple t
{
  int beta;
  int alp;
  int r;
};

t allData[M]=[<1,1,1>,<2,2,2>];

int maxr=max(m in M) allData[m].r;
       
dvar int y[M][1..maxr][N] in 0..1;

subject to
{
y[1][1][1]==2;
}

or use tuple sets instead since arrays with variable size are not allowed in OPL

regards