The phrasing of the second constraint is odd, in that it implies the value of a parameter is conditioned on the value of a variable. Are you trying to say that if any of t[i, j, j+1] ... t[i, j, j + l[i]] takes any value other than 1 then z[i,j] must be 0? If so, what is the domain of t[i, j, k] (binary, nonnegative integer, positive integer, real within specified bounds, ...)?
------------------------------
Paul Rubin
Professor Emeritus
Michigan State University
------------------------------
Original Message:
Sent: Sat June 19, 2021 02:43 AM
From: Adnan Pasha
Subject: Using sum and forall with a range consisting of a different input parameter
Thanks a lot for responding. I agree that for the first constraint, writing index like that was very naive on my part. I am sorry for that. I just changed the index and it worked. However, I don't think there is any problem with the indices of the second constraint. Is there any way to model that kind of a constraint in CPLEX.
Thanks in anticipation.
------------------------------
Adnan Pasha
Original Message:
Sent: Thu June 17, 2021 01:50 PM
From: Paul Rubin
Subject: Using sum and forall with a range consisting of a different input parameter
The problem is that the mathematical statement of the constraints is notational nonsense. You have the sum as j goes from j to whatever for all j in J. In other words, "j" means two entirely different things within the same constraint. Try changing the index of summation to something like j1 and see if that resolves the problem.
------------------------------
Paul Rubin
Professor Emeritus
Michigan State University
Original Message:
Sent: Thu June 17, 2021 08:09 AM
From: Adnan Pasha
Subject: Using sum and forall with a range consisting of a different input parameter
Hi,
I am new to Decision optimization and OPL. I am not able to model two constraints in cplex
I tried to model like this respectively, but it is not working
forall(i in B, j in S, c in C)
(sum(j in S: j in j..(j+l[i]-1)) X[j,c] - Y[i,c]) * z[i,j] >= 0;
forall(i in B, j in S, k in S: k in j+1..j+l[i])
t[i,j,k] <= z[i,j];
where z[i,j] is a binary decision variable
X[j,c], Y[i,c] and t[i,j,k] are input parameters
l[i] is also an integer input parameter
I shall be very thankful for any kind of help.
------------------------------
Adnan Pasha
------------------------------
#DecisionOptimization