a specific question on part of the AMPL model

29 views
Skip to first unread message

Sattar Ezzati

unread,
May 14, 2018, 10:33:34 AM5/14/18
to am...@googlegroups.com

Hello Robert,

 

I hope all is well with you, I have problems with my model and I don’t know why this happens,

 

The question is

I wonder does I write down the inventory constraints correctly in the AMPL? If yes, why the variables zSrt(hst) does not get value, sometimes?  

When I run the model with constraints (1-5) I have values for all variables except for zSrt(hst), and I don’t know why? (please run the model to see)

 

Also when I ignore only constraints 5, and add another constraints such as process variables (Zprc (jrt) the model gives me an infeasible solution, (please ignore constraint 5 and open constraints 6 and 6+1 and 7, and then run the model to see what I try to say.

 

I have attached a math. Version, corresponding code in AMPL and a small example to be able to run the model, could you please take a look on my model and let me know where I made mistakes.

 

Sorry, if I could not post my problem on Forum, because this model still is under development and does not spread out yet.  

 

Thanks a million for your help, guidance and time.

 

-----------------------------------------------------------------------

Sättar Ezzati

 

Université Laval

 

Département de Génie Mécanique

Pavillon Adrien-Pouliot, bureau #3357
1065, avenue de la Médecine
Université Laval, Québec (Québec) G1V 0A6, Canada

Téléphone |  +1 (418) 656 2131 #2414   

Mob.1 |  +1 (418) 271 1864 

Mob.2 |  +46 (700) 23 6574

P.Box | 6838

 

 

 

Capture.JPG
Data.txt
desB.xlsx
Model.txt
Mathematical model-form.docx

AMPL Google Group

unread,
May 15, 2018, 10:40:20 AM5/15/18
to Ampl Modeling Language
When you say that the variables zSrt[h,s,t] do not get a value, do you mean that they all have the value 0 in the optimal solution? It seems that could happen in an optimal solution; the zSrt[h,s,t] variables have a cost in the objective being minimized, and it appears that the constraints could be satisfied when they are all zero. If this is the situation, then you should check that the solver returned an "Optimal Solution" message, and if so, then you need to study your model more closely to determine why it is optimal to have all of those variables be zero.

In cases where the problem is not feasible, you also have to do some study to determine the reason. Here it may help to compute an irreducible infeasible subset (IIS) as explained in https://groups.google.com/d/msg/ampl/o8aNfwpoBAc/D46d6cfICQAJ. (This describes the IIS feature for CPLEX, but there are very similar features in Gurobi and Xpress.)

--
Robert Fourer
am...@googlegroups.com
{#HS:580534245-7658#}
--
You received this message because you are subscribed to the Google Groups "AMPL Modeling Language" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ampl+uns...@googlegroups.com.
To post to this group, send email to am...@googlegroups.com.
Visit this group at https://groups.google.com/group/ampl.
For more options, visit https://groups.google.com/d/optout.



Reply all
Reply to author
Forward
0 new messages