Figuring out Objective Function Zero Value Reason

Problems with modeling
Post Reply
Behnam47
User
User
Posts: 2
Joined: 2 years ago

Figuring out Objective Function Zero Value Reason

Post by Behnam47 »

Hi, I've been coding a Two Way supply chain Program for so long. I got to this place that my Objective Function returns a 0 Value with a upper Bound: 0 too.
my question is that is there any way or tool or option that can help me Figure out which Constraint(s) are making the problem to have no feasible area?
it's a maximization problem and i've been using Baron Solver. I will attach my code here.
I really appreciate if anyone can help.
Small-Model03.gms
(12.94 KiB) Downloaded 159 times
abhosekar
Moderator
Moderator
Posts: 295
Joined: 3 years ago

Re: Figuring out Objective Function Zero Value Reason

Post by abhosekar »

It is wrong to say that it is a problem and that there is no feasible region. You are getting a feasible solution and a proven optimal solution. Now a good way to tackle such problems is based on your understanding of the problem. Do you see some of the quantities that are 0 (and should not be zero)? Do you have penalty terms that drive optimization in such a direction (for example, if a transport cost is too high compared to sales revenue, optimization can tell you that it is best not to transport anything at all). I see a lot of negative terms in the objective function and only one positive term.
sum((k,c1,t,s),pi(s)*(QKL(k,c1,t,s)*prl(c1,t,s)))

However, QKL appears also in the negative term
(sum((k,c1,t),QKL(k,c1,t,s)*tcc(k,c1)))

If the combined coefficient of QKL is negative (and since QKL is a positive variable), the optimal value for this should be 0.

If you remove all negative terms in the objective funciton and just keep sum((k,c1,t,s),pi(s)*(QKL(k,c1,t,s)*prl(c1,t,s))) you get a positive solution. I would suggest checking parameters and tables.

Hope this gives you something to think about your formulation.

- Atharv
Post Reply