Flat objective function

GAMS
1

Dear all,

I got a return from CONOPT solver as follows:

** Optimal solution. Reduced gradient less than tolerance.


The explanation from CONOPT manual is that the value of the objective function is very accurate while the values of the variables are less accurate due to a flat objective function in the interior of the feasible area. If I am interest in variables other than objective function, is this result reliable? Is there any way to improve the solution?


Many thanks,

Qiaomin

–
To unsubscribe from this group and stop receiving emails from it, send an email to gamsworld+unsubscribe@googlegroups.com.
To post to this group, send email to gamsworld@googlegroups.com.
Visit this group at http://groups.google.com/group/gamsworld.
For more options, visit https://groups.google.com/d/optout.

2

Qiaomin:



You can think of the following very simple model:



minimize sqr(x)+1



The solution is x=0 with objective 1. A small change in x to say 1.e-8 will give an objective of 1+1.e-16 which internally in the computer will be rounded to 1. If CONOPT comes to the point 1.e-8 it may not be able to improve it because the possible change in objective is so small, less than machine precision.



In general, the objective will be OK to 10-12 decimals; the variables to around 6-7 decimals.



Regards



Arne

Arne Stolbjerg Drud

ARKI Consulting & Development A/S

Bagsvaerdvej 246A, DK-2880 Bagsvaerd, Denmark

Phone: (+45) 44 49 03 23, Fax: (+45) 44 49 03 33, email: adrud@arki.dk



From: gamsworld@googlegroups.com [mailto:gamsworld@googlegroups.com] On Behalf Of Qiaomin Li
Sent: Friday, April 11, 2014 12:06 PM
To: gamsworld@googlegroups.com
Subject: Flat objective function



Dear all,



I got a return from CONOPT solver as follows:



** Optimal solution. Reduced gradient less than tolerance.



The explanation from CONOPT manual is that the value of the objective function is very accurate while the values of the variables are

less accurate due to a flat objective function in the interior of the feasible area. If I am interest in variables other than objective function, is this result reliable? Is there any way to improve the solution?



Many thanks,

Qiaomin

–
To unsubscribe from this group and stop receiving emails from it, send an email to gamsworld+unsubscribe@googlegroups.com.
To post to this group, send email to gamsworld@googlegroups.com.
Visit this group at http://groups.google.com/group/gamsworld.
For more options, visit https://groups.google.com/d/optout.

–
To unsubscribe from this group and stop receiving emails from it, send an email to gamsworld+unsubscribe@googlegroups.com.
To post to this group, send email to gamsworld@googlegroups.com.
Visit this group at http://groups.google.com/group/gamsworld.
For more options, visit https://groups.google.com/d/optout.