Originally posted by: ikucukkoc

 

Hello

I have defined required data, decision variables, and objective function. However, I have a problem to code the constraint attached to this post.

{int} ProdCycles = {1, 2, 3, 4, 5, 6};

{int} Queues = {1, 2, 3, 4};

int TotalPhi = 6;

tuple linemodel{
  int h;
  string j;
}

{linemodel} LineModels = {<h, j> | <h, j, i> in LineModelTasks}; //LineModelTasks has defined in the code

tuple propdemand{
  key string j;
  float d;
};

propdemand PropDemandsArray[i in 1..card(PropDemands)] = item(PropDemands, i-1);

 

 

My question is: I defined "<h, j> in LineModels" for forall section but I also need to sum all "h in Lines". Is that a proper representation for the attached constraint? Or should I delete "h in Lines"?

The next question is the program does not accept "PropDemandsArray[<j,i>". How should I modify it?

Many thanks in advance.

Best Regards

 

#DecisionOptimization
#OPLusingCPLEXOptimizer

Hi,

 

since you have 3 sigmas in your formula, you should keep 3 elements in the sum.

PropDemandsArray is an array and its index should be an int : i in 1..card(PropDemands);

but in

T[ phi, <h, j>, q ] == TotalPhi * PropDemandsArray[<j,i>];

you use a tuple as index, which is not good.

Plus PropDemandsArray[<j,i>] is type tuple and cannot be multiplied.

Can you attach your .mod and .dat so someone can have a look ?

 

Regards

#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: ikucukkoc

 

Hi Alex

 

Many thanks for your reply.

 

I am attaching whole .mod and .dat files. The files may seem a bit complex but I need your help to make sure that whether my code is correct to define the Objective Function (3.9) and Constraints (3.13, 3.14, and 3.15) in the attached ikkoc2.png file.

 

Your help is much appreciated.

 

Many thanks in advance.

 

Best Regards

 

#DecisionOptimization
#OPLusingCPLEXOptimizer

Hi,

turning

forall( <h, j> in LineModels )
    ct314: //Model Demand Satisfaction Constraint
      sum( phi in ProdCycles, /*h in Lines,*/ q in Queues )
        T[ phi, <h, j>, q ] == TotalPhi * PropDemandsArray[<j,i>];

 

into
        
        
 forall( <h, j> in LineModels )
    ct314: //Model Demand Satisfaction Constraint
      sum( phi in ProdCycles, /*h in Lines,*/ q in Queues )
        T[ phi, <h, j>, q ] == TotalPhi * PropDemandsArray[ord(PropDemands,<j,h>)].d;

 

helps.

I let you go on like this and fix your model

 

Regards

#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: ikucukkoc

 

Many thanks for the reply. Sorry, there should not be comments before and after h in lines. So the sum section would be "sum( phi in ProdCycles, h in Lines, q in Queues )".   So, should i keep "h in lines" in sum? is it correct in this format?

 

forall( <h, j> in LineModels )

    ct314: //Model Demand Satisfaction Constraint

      sum( phi in ProdCycles, h in Lines, q in Queues )

        T[ phi, <h, j>, q ] == TotalPhi * PropDemandsArray[ord(PropDemands,<j,h>)].d

 

Is that correct now without comments? I am asking because I also have "h" as "<h, j> in LineModels". So,does "h in Lines" duplicate "h"?

Really appreciate for your help. I solved examples but cannot find solutions for such specific problems. And your replies really helped me a lot. And I am sure that these conversations will also be helpful for other forum followers.

 

Best Regards

 

#DecisionOptimization
#OPLusingCPLEXOptimizer

forall( <h, j> in LineModels )
    ct314: //Model Demand Satisfaction Constraint
      sum( phi in ProdCycles, /*h in Lines,*/ q in Queues )
        T[ phi, <h, j>, q ] == TotalPhi  ; //* PropDemandsArray[<j,i>];

  forall( <h, j, i> in LineModelTasks, phi in ProdCycles)
    ct315: //Task Occurence Constraint
      sum( h2 in Lines, q in Queues, x in Sides, k in Workstations )
        T[ phi, <h,j>, q ] * Y[ phi, <h, j, i>, k, x ] <= 1;

 

works and you can debug from that stage.

Becareful : you used h in <h, j, i> in LineModelTasks and sum( h in Lines)

is is really the same h ?

 

Regards

#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: ikucukkoc

Thanks for the reply. 

 

1) So you do not define h second time in ct314 because you are keeping comment command /*h in Lines,*/ ??

 

2) In ct315, you define a h2 but you do not use it, is that right?

 

3) Actually I do not have to use it like  "<h, j, i> in LineModelTasks" but "i"s are defined for each j, and "j"s are defined for each "h". That is why i use "<h, j, i> in LineModelTasks" Is there another way for such expression please?

 

Regards

 

#DecisionOptimization
#OPLusingCPLEXOptimizer