Path with variable inside

GAMS
1

Hi!

I would like to change the path within a for loop.

The path looks like: C:\Data\Instanz Num 001.xlsm

and the 3 numbers at the end of the path are like a variable which must be counted up for each loop.

How can I write this path?



Thank you in advance :wink:

2

What do you want to do with these paths? Write a put file or unload a GDX file there? These can be arranged with put_utilities. See https://www.gams.com/latest/docs/UG_Put.html#UG_Put_PutUtil. It seems that you want to do something with the Excel files. Again put_utility ‘exec’ allows you to call programs with arguments build at execution time.

-Michael

3

The paths are needed for loading some data.

The code should look like this, while the bold highlighted Nummer in the path should be instead the intN value

for (intN=1 to 100,

$call gdxxrw “C:\Users\manue\Daten\i10 Instanz 001.xlsm” Squeeze=N @C:\Users\manue\Daten\Index_det.txt
$gdxin “C:\Users\manue\Daten\i10 Instanz 001.gdx”
$load i k
$gdxin

…some futher arguments… ;

);

Therefore my question is how I can add it to the path.
I know that in VBA for example I could write the path like: “C:\Users\manue\Daten\i10 Instanz” &intN & “.xlsm”
Is something similar possible in GAMS?

4

This does not work. In the other post I was pointing you to a chapter understanding compile- and runtime in GAMS. All complile time statements (like $gdxin/$load are done at compile time) while others like loop(…) are done at execution time. So you cannot mix and match them. Is sounds like that you have a model that you want to run for each instance in a loop. Separate the two:

* This mymodel.gms is a model for one instance
$call gdxxrw "%fn%.xlsm" output="%fn%.gdx" Squeeze=N @C:\Users\manue\Daten\Index_det.txt
$gdxin "%fn.gdx%"
$load i k
$gdxin

* The model 
...




* The scenario driver
set i / 001*100 /;
loop(i,
   put_utility 'exec' / 'gams mymodel --"C:\Users\manue\Daten\i10 Instanz ' t.tl:0 '" lo=%gams.lo% o=mymodel' t.tl:0 '.lst';
);

-Michael

5

Dear Mr Michael,

Can you help me by answering this question. I have a problem which has quadratic objective function and linear constraints. All variables are continuous. I used BARON solver with model type QCP. IS the solution obtained global optimal? Do I have to adjust any option in the solver? Or any local optimal will be global as this is a convex problem So I guarantee to have the global optimal solution reached.

Thanks a lot.

6

Thank you very much!

I do have some questions about the code you wrote.

  1. The %gams.lo% in lo=%gams.lo%. What does it do?
  2. Are my changes correct? I insert i.tl instead of t.tl and “fn=” infront of the path.
  3. Do I have to declare fn somewhere?