I am confused about this situation: I have a binary variable \(x_{ij}\) and an integer variable \(y_j\). Also, if \(x_{ij}=1\) then \(y_j\) decreases by \(1\), i.e., \(y_j\gets y_j1\). Here, \(y_j=c_j\) at the start of the otpimization process. I don't know if I can express this constraint as a linear one? Can I juste write: \(y_j=y_jx_{ij}\)? Doesn't this forces \(x_{ij}=0\) all times? asked 17 Jan '18, 18:21 zbir 
Doesn't \(y_j\) have a starting value that is constant? If that is \(c_j\), could'nt it be written as \(y_j=c_j\sum_i x_{ij}\) ? Maybe I did not understand the question right. answered 17 Jan '18, 18:26 Naveen Divak... Yes, \(y_j=c_j\) at the start. See the updated question.
(17 Jan '18, 18:29)
zbir
then ((y_j=c_j\sum_{i}x_{ij})) should work
(17 Jan '18, 18:38)
Naveen Divak...
Was my answer helpful? Or did I still understand your question wrong?
(18 Jan '18, 01:49)
Naveen Divak...
I think you did understand the question correctly but which one is the right answer? Is it \(y_j=c_jx_{ij}\) or \(y_j=c_j\sum_i x_{ij}\)?
(18 Jan '18, 10:18)
zbir
(y_j=c_j\sum_i x_{ij})
(18 Jan '18, 10:21)
Naveen Divak...
