Dear all,
I need to solve a minimization problem with 300 variables that form 1000 combinations in pairs of two. Not all possible combinations are made and sometimes the variables can combine themselves two times. Additionally:
1- There are is also a parameter "c" that depends on the variables present in the combination, regardless of the order,
2- And two parameters "a" and "b" that depend on the order of the variables.
As an example:
(a*x1+b*x2 - c12 )^2 + (a*x2+b*x2 - c22 )^2 + (a*x2+b*x1 - c21 )^2
where c21=c12
After reading the provided materials my main question is: how can I set up my model and data in such a way that AMPL understands that the second variable appearing corresponds to the first variable in some other combinations? That is, how can I show that a*x1 in the first combination above and b*x1 in the third combination above correspond to the same variable?