Originally posted by: SystemAdmin

Hi,

I've been trying to implement a model in CPLEX for a while now. But it seems I just can't get the indices right...
I get the error msg that "range is not available for scripting". Whatever that means...
Basically what I want to model is this: There is a variable called "stages" It has the index j and j runs from 1 to the number of stages (in my case that's 3, I defined it as Nbstages). In one of the constraints I need to access j-1, so I only want to have this constraint for j in 2..Nbstages (as j is only defined from 1 to Nbstages and j-1 would be 0 for j=1).

I would be very very grateful for some help on this topic. This is the code I came up with:

int Nbstages = ...;
range stages = 1..Nbstages;
int LeadTime stages=...;
int MaxServiceTime stages=...;

float safetyfactor stages=...;
float StandardDeviation stages=...;
int MeanDmd stages=...;
float StageCost stages=...;
float HoldingCostRate stages=...;
float HoldingCost stages=...;;

dvar int+ OutboundServiceTime stages;
dvar int+ InboundServiceTime stages;

minimize
sum (j in stages)
(HoldingCost[j]*(MeanDmd[j]*(InboundServiceTime[j]+LeadTime[j]-OutboundServiceTime[j]))
+(safetyfactor[j]*StandardDeviation[j]*/*sqrt*/(InboundServiceTime[j]+LeadTime[j]-OutboundServiceTime[j])));
subject to {
forall (j in stages)
OutboundServiceTime [j]-InboundServiceTime[j] <= LeadTime[j];
forall (j in stages)
InboundServiceTime [j]-OutboundServiceTimej-1 >= 0;
//forall (j in stages)
OutboundServiceTime[0]==0;
forall (j in stages)
OutboundServiceTime [j]<= MaxServiceTime [j];
forall (j in stages)
InboundServiceTime [j]>=0;
forall (j in stages)
OutboundServiceTime [j]>=0;
};

Thanks a lot,
Carolin
#DecisionOptimization
#OPLusingCPLEXOptimizer

Hi,

could you attach your model and data file?

regards
#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: SystemAdmin

Of course. Thanks so much for taking a look!

Some of the data in the data file is not the real data I will work with later on. I still have to figure out a way how to aggregate some of my initial data to be able to use it. I just entered some placeholders to see if I could get the model to work.

Best,
carolin
#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: SystemAdmin

...and here's the data file.
#DecisionOptimization
#OPLusingCPLEXOptimizer

Hi

I rewrote your constraints as

subject to {
    forall (j in stages)
    OutboundServiceTime [j]-InboundServiceTime[j] <= LeadTime[j];
  forall (j in stages:j!=1)
    InboundServiceTime [j]-OutboundServiceTime[j-1] >= 0;
  //forall (j in stages)
    //OutboundServiceTime[0]==0;
  forall (j in stages)
    OutboundServiceTime [j]<= MaxServiceTime [j];
  forall (j in stages)
    InboundServiceTime [j]>=0;
  forall (j in stages)
    OutboundServiceTime [j]>=0;
    
  };

regards
#DecisionOptimization
#OPLusingCPLEXOptimizer

PS:

You can use range in script as in

execute
  {
   for(var i in stages) writeln(i); 
  }

#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: SystemAdmin

Thank you so much for your help!
#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: SystemAdmin

Sorry to bother you again, but I have another question. My objective function contains a square root of the sum of some variables. However, this obviously doesn't work that easily. In the beginning I thought it was because of the indexing issue. But it's not.

This is the objective function:

minimize
sum (j in stages)
(HoldingCost[j]*safetyfactor*StandardDeviation*sqrt(InboundServiceTime[j]+LeadTime[j]-OutboundServiceTime[j]));

I tried to find out the confinements for using the sqrt function. The only thing I found was that Cplex can only extract roots from float values. I changed the integer variables to floats but that didn't help.

Is there another constraint for the usage of square roots in an objective function?

Thanks,
Carolin
#DecisionOptimization
#OPLusingCPLEXOptimizer

Hi,

since your variables are integer, why don't use use Constraint Programming?

In the code, you simply need to do using CP;

using CP;
 
 
//{string} stages=...;
 
int Nbstages = ...;
range stages = 1..Nbstages;

Regards
#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: SystemAdmin

Hi,

thanks again. That was a lot easier than expected. :) The model works perfectly now.
Just one more question. If I wanted to make a restriction on the possible values of a variable, how would I do that?

For example I'd like to say that OutboundServiceTime [j] can either be InboundServiceTime [j]+LeadTime[j] or zero. I tried modelling it like this, but it's not working.

forall (j in stages)
OutboundServiceTime [j] == [InboundServiceTime[j]+LeadTimej] 0;

The only information I could find in the documentation is that the domain of variables is denoted with square brackets...

Thanks,
Carolin
#DecisionOptimization
#OPLusingCPLEXOptimizer

Hi

why don't you do

(OutboundServiceTime [j] == InboundServiceTime [j]+LeadTime[j] )
||
(OutboundServiceTime [j] == 0)

regards
#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: SystemAdmin

Thanks so much. It worked. :)

Unfortunately I have a new problem. I tried to implement a second model but again I have problems with the index ranges... This model is a bit more complicated as it contains two sets of indexes and I have to sum up over both of them.
Also, I sometimes want to include variables in the index ranges. So for example I might want to sum up over all preiods starting with the current period and ending with the period that equals the transportation lead time of the stage. I don't really know how to do this in OPL...

I hope you can help me again. I would be so grateful!
Thanks,
Carolin
#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: SystemAdmin

data file
#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: SystemAdmin

new question added
#DecisionOptimization
#OPLusingCPLEXOptimizer