absolute gap

GAMS
1

hi

i have modeled a LRP with objective function includes two components such as routing cost and location cost in gams, but for large size problem (specially in number of customer) location cost doesnt work properly.In act when i increase unit cost for opening depot it doesnt try to open less number of depots and its not correct absolutely .I have added option optca=0 but it was not useful to solve my problem too.Also report shows absolute gap=129960

can any one help me on this issue??

–
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

Hello ??

Have you tried optcr=0?

Regards

Fede

On Thu, Jan 29, 2015 at 4:05 PM, saba wrote:

Hi dear friends

How can i reduce absolute gap in solving MIP model?? In one of my reprts in gams I have MIP solution =5000 and Best solution=2500 and absolute gap=2500 although I add option optca=0 but it has not effect for reducttion named gap too. Can any one help me on this issue??

regards,


\

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.

3

Hi
dear Fedperea.

no I have not!

In fact, I defined optcr=0.05 and optca=0 . Is it incorrect??


On Thursday, 29 January 2015 19:35:58 UTC+4:30, saba wrote:

Hi dear friends

How can i reduce absolute gap in solving MIP model?? In one of my reprts in gams I have MIP solution =5000 and Best solution=2500 and absolute gap=2500 although I add option optca=0 but it has not effect for reducttion named gap too. Can any one help me on this issue??

regards,


\

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.

4

I would try optcr =0

On Fri, Jan 30, 2015 at 6:08 AM, saba wrote:

Hi
dear Fedperea.

no I have not!

In fact, I defined optcr=0.05 and optca=0 . Is it incorrect??


On Thursday, 29 January 2015 19:35:58 UTC+4:30, saba wrote:

Hi dear friends

How can i reduce absolute gap in solving MIP model?? In one of my reprts in gams I have MIP solution =5000 and Best solution=2500 and absolute gap=2500 although I add option optca=0 but it has not effect for reducttion named gap too. Can any one help me on this issue??

regards,


\

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.

5

On Thursday, 29 January 2015 18:35:58 UTC+3:30, saba wrote:

Hi dear friends

Hi


Thanks alot.
I did it ,but it had not effect

regards,




How can i reduce absolute gap in solving MIP model?? In one of my reprts in gams I have MIP solution =5000 and Best solution=2500 and absolute gap=2500 although I add option optca=0 but it has not effect for reducttion named gap too. Can any one help me on this issue??

regards,


\

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.

6

Dear Saba

If you are using the CPLEX solver, try the following CPLEX options

epagap absolute stopping tolerance
epgap relative stopping tolerance

Regards;

Hossein

On Sat, Jan 31, 2015 at 2:30 PM, saba wrote:


On Thursday, 29 January 2015 18:35:58 UTC+3:30, saba wrote:

Hi dear friends

Hi


Thanks alot.
I did it ,but it had not effect

regards,




How can i reduce absolute gap in solving MIP model?? In one of my reprts in gams I have MIP solution =5000 and Best solution=2500 and absolute gap=2500 although I add option optca=0 but it has not effect for reducttion named gap too. Can any one help me on this issue??

regards,


\

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.

7

hi
thanks dear hossein

but i have two questions : how i should code(write) epgap and epagap in model? and if i use epgap and epagap, its not necessary to use optcr and optca?!

regards

On Thursday, 29 January 2015 19:35:58 UTC+4:30, saba wrote:

Hi dear friends

How can i reduce absolute gap in solving MIP model?? In one of my reprts in gams I have MIP solution =5000 and Best solution=2500 and absolute gap=2500 although I add option optca=0 but it has not effect for reducttion named gap too. Can any one help me on this issue??

regards,


\

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.

8

Dear Saba



First, force Cplex to read the option file:

ModelName.OptFile = 1;

(Instructs Cplex to read the option file. The name of the option file is cplex.opt.)

Then, make the option file and put your desired options. You can use the following code. Just put your model name instead of " ModelName ".


ModelName.Optfile =1;

file opts cplex option file/ cplex.opt /;

putclose opts /‘epgap=0’ /‘epagap=0’;


You can study CPLEX manual for more details.



Regards;



Hossein

On Sun, Feb 1, 2015 at 8:05 PM, saba wrote:

hi
thanks dear hossein

but i have two questions : how i should code(write) epgap and epagap in model? and if i use epgap and epagap, its not necessary to use optcr and optca?!

regards

On Thursday, 29 January 2015 19:35:58 UTC+4:30, saba wrote:

Hi dear friends

How can i reduce absolute gap in solving MIP model?? In one of my reprts in gams I have MIP solution =5000 and Best solution=2500 and absolute gap=2500 although I add option optca=0 but it has not effect for reducttion named gap too. Can any one help me on this issue??

regards,


\

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.

9

These Cplex options override GAMS option. So, these option always applies whether GAMS option (optcr and optca) write or not provided that reading option file is enabled (ModelName.OptFile = 1;). Read Cplex manual:



epagap (real)

Absolute tolerance on the gap between the best integer objective and the objective of the best node remaining. When

the value falls below the value of the epagap setting, the optimization is stopped. This option overrides GAMS OptCA

which provides its initial value.

(default = GAMS OptCA)



epgap (real)

Relative tolerance on the gap between the best integer objective and the objective of the best node remaining. When

the value falls below the value of the epgap setting, the mixed integer optimization is stopped. Note the difference in

the Cplex definition of the relative tolerance with the GAMS definition. This option overrides GAMS OptCR which

provides its initial value.

Range: [0,1]

(default = GAMS OptCR)

On Mon, Feb 2, 2015 at 10:11 AM, Hossein Sharifzadeh wrote:

Dear Saba



First, force Cplex to read the option file:

ModelName.OptFile = 1;

(Instructs Cplex to read the option file. The name of the option file is cplex.opt.)

Then, make the option file and put your desired options. You can use the following code. Just put your model name instead of " ModelName ".


ModelName.Optfile =1;

file opts cplex option file/ cplex.opt /;

putclose opts /‘epgap=0’ /‘epagap=0’;


You can study CPLEX manual for more details.



Regards;



Hossein

On Sun, Feb 1, 2015 at 8:05 PM, saba wrote:

hi
thanks dear hossein

but i have two questions : how i should code(write) epgap and epagap in model? and if i use epgap and epagap, its not necessary to use optcr and optca?!

regards

On Thursday, 29 January 2015 19:35:58 UTC+4:30, saba wrote:

Hi dear friends

How can i reduce absolute gap in solving MIP model?? In one of my reprts in gams I have MIP solution =5000 and Best solution=2500 and absolute gap=2500 although I add option optca=0 but it has not effect for reducttion named gap too. Can any one help me on this issue??

regards,


\

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.

10

hi dear hossein

you say right but my question is different with your answer!In fact I askes you say me whether two sentences such as option OptcR=0.05 and
option OptcA=0 before “solve …using mip minimizing…” are enough or I should use epgap and epagap?. and my another question is how i can write them??

regards

On Thursday, 29 January 2015 19:35:58 UTC+4:30, saba wrote:

Hi dear friends

How can i reduce absolute gap in solving MIP model?? In one of my reprts in gams I have MIP solution =5000 and Best solution=2500 and absolute gap=2500 although I add option optca=0 but it has not effect for reducttion named gap too. Can any one help me on this issue??

regards,


\

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.