Originally posted by: rapherd

Dear all,

I'm not using ILog such often. Therefore, I want to ask if it's possible to generate an example DAT-File based on a given model (mod-File) definition? My intension is based on some generators that are available in OOP languages for generating Java classes based on given XSD schema files. I think it's very useful even if the data are not semantically useful but it saves a lot of time and pain :).

Thank you for your help

Regards

Raphael

The data section in a .mod already is very similar to a .dat file. Can't you just copy-paste that and make the required modifications manually? Or are you looking for an automatic way to do that?

Originally posted by: rapherd

Hello,

hm, I'm not sure if you understand my question correct. Therefore, I decided to give an example. Let's say my model (model.mod) looks like:

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

///////////////// Sets and indices /////////////////
 {string} P = ...;                // all available products
 {string} FE = ...;                // FE facilities
 {string} BE = ...;                // BE facilities
 {string} DB = ...;                // DB facilities
 {string} DC = ...;                // DC facilities
 int MP = ...;                    // maximum nuber of periods
 range T = 1..MP;                // Time Range 

///////////////// Parameters ///////////////// 
 int h_FE[FE][P] = ...;            // holding cost for FE facilities
 int b_FE[FE][P] = ...;            // backlog cost for FE facilities
 int w_FE[FE][P] = ...;            // WIP cost for FE facilities

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

and in OPL-Studio I'm able to define a set of data in (data.dat) that are bound to this variables. Let's assume this:

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

P = {"part_1"};
FE = {"FE1" "FE2"};                // FE facilities
BE = {"BE1" "BE2"};                // BE facilities
DB = {"DB1" "DB2"};                // DB facilities
DC = {"DC1" "DC2"};                // DC facilities
MP = 5;                            // maximum nuber of periods

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

So, what I'm looking for is a way of generating this data.dat automatically based on my model.mod file. The semantical meaning of the content in data.dat is irrelevant. Important is to have a set of values that matches the syntactical definitions from the model.mod file.

The shown example is simple, but if you have multi-dimensional arrays it's a pain to format the data.dat correct.

Thank you for any hint

Regards

Raphael

Hi,

first you could use any language to build the .dat file, even scripting in OPL.

You have an example in Python at

https://www.ibm.com/developerworks/community/forums/html/topic?id=0b6cacbe-4dda-4da9-9282-f527c3464f47

and in R at

https://www.ibm.com/developerworks/community/forums/html/topic?id=1ba857cb-6d60-44cb-b645-1189b25c988c&ps=25

But this requires a bit of coding.

You may also try something like

///////////////// Sets and indices /////////////////
 {string} P;;                // all available products
 {string} FE;;                // FE facilities
 {string} BE;;                // BE facilities
 {string} DB;;                // DB facilities
 {string} DC ;                // DC facilities
 int MP ;                    // maximum nuber of periods
 range T = 1..MP;                // Time Range

///////////////// Parameters /////////////////
 int h_FE[FE][P] ;            // holding cost for FE facilities
 int b_FE[FE][P];            // backlog cost for FE facilities
 int w_FE[FE][P];            // WIP cost for FE facilities
 
 execute
 {
 thisOplModel.settings.forceUsage=true;
 }
 
 
 main
 {
 thisOplModel.generate();
 
 writeln(thisOplModel.printInternalData()) ;
 }

which will give you

MP = 0;
// RANGE_INT T [0 ,-1]
DC = {};
DB = {};
BE = {};

that you could send to a file with IloOplOutputFile

regards

Originally posted by: rapherd

Hey,

thanks for your idea but thats not very useful. My output of your snippet is quite similar:

FE = {};
P = {};
w_FE = [];
b_FE = [];
h_FE = [];
MP = 0;
// RANGE_INT T [0 ,-1]
DC = {};
DB = {};
BE = {};

So, what I see is, that for example, w_FE is an array because of the braces but the reason of my initial question is, how is the format of this array? I'm talking about something like

p=

[[23.64 23.64 23.64 23.64 23.64 23.64]

 [110.16 110.16 110.16 110.16 110.16 110.16]

 [110.16 110.16 110.16 110.16 110.16 110.16]

 [110.16 110.16 110.16 110.16 110.16 110.16]];

which is an 2D- int array int p[Size1][Size2].

But if I understand that correct, there is no really useful way to extract that data?

Thanks for any kind of further help

Raphael

Well, you can always write a small program that loads a model and then inspects it. For example

  public static void main(String[] args) throws IloException {
    IloOplFactory factory = new IloOplFactory();
    for (final String mod : args) {
      final IloOplRunConfiguration rc = factory.createOplRunConfiguration(mod, new String[0]);
      final IloOplModel model = rc.getOplModel();
      model.generate();
      for (Iterator<?> it = model.getElementIterator(); it.hasNext(); /* nothing */) {
        final IloOplElement element = (IloOplElement)it.next();
        // Switch over the different element types here
        ...
      }
    }
  }

From the innermost loop you can print the information you want. The reference documentation should help you in figuring out, which element types exist etc.