Originally posted by: ChWeil

Hi guys, 

I have a model in which one part of the objective function is formulated as follows:

(sum j in jobs, v in variations) q(i,v) * (C(j)-dd(j))

with

float dd(j)

float q(i,v)

dvar float+ C(j)

dvar boolean z1(j,v), z2(j,v), z(j,v)

My problem is that the value of parameter q(i,v) depends on the deviation of C(j) and dd(j)

Therefore, I tried to model a constraint like that:

forall (v in V, j in J) (l(v)*dd(j) <=C(j)) => z1(j,v)==1, where l(v) state a lower bound 

forall (v in V, j in J) (C(j) <= u(v)*dd(j)) => z2(j,v) ==1, where u(v) state an upper bound

forall (v in V, j in J) z(j,v) + 1 == z1(j,v)+z2(j,v)

Unfortunately the model only gives z1(j,v) and z2(j,v) ==1 for all values of C(j) which is not true.

Another idea was to model a decision variable like dvar boolean x(j,v) -> 1, if l(j,v)*dd(j) <= C(j) <= u(j,v)*dd(j), 0 else but I got an error that C(j) is not allowed.

Can someone help me with that and know what I am doing wrong? 

Thank you 

Best

CW

Hi

for arrays you should use brackets instead of parenthesis.

see

https://www.linkedin.com/pulse/making-decision-optimization-simple-alex-fleischer

regards

Originally posted by: ChWeil

Thanks for your reply.

Actually I wrote everything in the form of e.g. z1[j] [v] in my model so unfortunately that is not the fault.

Do you have another suggestion what could be wrong? 

Hi

can you share .mod and .dat so that other users could help you ?

regards

Originally posted by: ChWeil

Hi, 

Yes of course. Find my project in the attachment.

I hope it gives a better understanding for the problem.

Originally posted by: ChWeil

Hi, 

Yes of course. I hope that can help. 

Best

CW