Originally posted by: Cplex

Hey guys,

I have a problem modelling a robust recoverable warehouse location problem. There's a while loop which contains the objective. It has to look more or less like this:

 

h == 0;

while(h <= 20){  
        FixC = sum (m in M) c[m]*x[m];
        VarC = sum (m in M,n in N) d[m][n]*y[m][n];
        RC = sum (n in N,k in K) v[n][k]*h*(1+p[k]);      

        costs = FixC + VarC + RC;
        minimize costs;
        h == h+1;}

 

So the costs should be minimized until h reaches a value of 20. The value of h influences the recoverable costs.

It would be nice if anyone could try to help me with this issue.

 

Thanks in advance!

 

#DecisionOptimization
#OPLusingCPLEXOptimizer

Hi,

why do not you adapt what was given at https://www.ibm.com/developerworks/community/forums/html/topic?id=fbea4b7b-6ed7-4fb8-a890-45df63b39222

?

You would add something like

dexpr RC[scenario in scenarii] = sum (n in N,k in K)
    scenario*(1+p[k]); 

and then the objective would be rewritten from

minimize sum(scenario in scenarii)
    (TotalFixedCost[scenario] + TotalSupplyCost[scenario]);

to

minimize sum(scenario in scenarii)
    (TotalFixedCost[scenario] + TotalSupplyCost[scenario]+RC[scenario]);

regards

 

 

#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: Kamran_Sr

Hi,

I have a model that I need to solve first, and then do a flow control to improve the solution. However, I got an issue with a while loop in this flow control. What I need to do is after solving the problem if an specific decision variable (SlackCap[i]) got a positive value, I have to change the data regarding to a set, which is the index of different decision variables, and rerun the model again until that decision variable get 0 value for all its elements. I have used a while loop to do this. The issue is, although incorporated a break statement (if (feasblity[i] = 0)  break;) for the loop, the model still keeps running even all elements of SlackCap[i] equals to 0.

My script code is as follows:

main {

 

var source = new IloOplModelSource("Feasiblity Restoration.mod");

var cplex = new IloCplex();

var def = new IloOplModelDefinition(source);

var model = new IloOplModel(def,cplex);

var data = new IloOplDataElements();

var k = 0;

 

model.addDataSource(data);

data.DataExcelFile = "InputData3.xlsx";

data.NbdryportsString="dryports";

data.fixedcostString = "fixedcost";

data.allocationcostString = "allocationcost";

data.coiString = "coim";

data.ceoiString = "ceoim";

data.bijString = "bijm";

data.beijString = "beijm";

data.li1String = "leaseI1";

data.li2String = "leaseI2";

data.viString = "vi";

data.hiString = "hi";

data.CapiString = "Capi";

data.LTString = "LTi";

data.RString = "Ri";

data.InversePhiNString = "InversePhiN";

data.InverseINString = "InverseIN";

data.InitialIString = "InitialI";

 

var Data = new IloOplDataSource("Feasiblity Restoration.dat");

model.addDataSource(Data);

model.generate();

 

if(cplex.solve()){

 

model.postProcess();

writeln("\t cost :", cplex.getObjValue());

writeln("\t SlackCap = ",model.SlackCap);

writeln("\t Location = ",model.X);

writeln("\t Allocation = ",model.Y);

 

var Infeasblity = new Array(model.dryports);

for(var i in model.dryports){

Infeasblity[i] = model.SlackCap[i].solutionValue;}

 

var dryports = model.dryports}

 

while(1){

for (var i in dryports){

if (Infeasblity[i]>0){

 

k++;

var model2 = new IloOplModel(def,cplex);

var data3 = new IloOplDataElements();

 

model2.addDataSource(data3);

data3.DataExcelFile = "Restoration.xlsx";

data3.NbdryportsString = "dryports"+k;

data3.fixedcostString = "fixedcost"+k;

data3.allocationcostString = "allocationcost"+k;

data3.coiString = "coim"+k;

data3.ceoiString = "ceoim"+k;

data3.bijString = "bijm"+k;

data3.beijString = "beijm"+k;

data3.li1String = "leaseI1"+k;

data3.li2String = "leaseI2"+k;

data3.viString = "_vi"+k;

data3.hiString = "_hi"+k;

data3.CapiString = "Capi"+k;

data3.LTString = "_LTi"+k;

data3.RString = "_Ri"+k;

data3.InversePhiNString = "InversePhiN"+k;

data3.InverseINString = "InverseIN"+k;

data3.InitialIString = "InitialI"+k;

 

var data4 = new IloOplDataSource("Feasiblity Restoration.dat");

model2.addDataSource(data4);

model2.generate();

 

if(cplex.solve()) {

model2.postProcess();

writeln("\t New cost"+k+" :", cplex.getObjValue());

writeln("\t New SlackCap"+k+" = ",model2.SlackCap);

writeln("\t New Location"+k+" = ",model2.X);

writeln("\t New Allocation"+k+" = ",model2.Y);

 

var dryports = model2.dryports;

for(var i in model2.dryports){

Infeasblity[i] = model2.SlackCap[i].solutionValue;}

 

var feasblity = new Array ();

for(var i in model2.dryports){

if (model2.SlackCap[i].solutionValue == 0)

feasblity[i] = model2.SlackCap[i].solutionValue}

 

}

}

else if (feasblity[i] = 0)

break;

}

if (feasblity[i] = 0)

break;

}

writeln("\t Final cost:", cplex.getObjValue());

writeln("\t Final SlackCap = ",model2.SlackCap);

writeln("\t Final Location = ",model2.X);

writeln("\t Final Allocation = ",model2.Y);

}

I would be grateful if someone can help me with this issue.

 

Regards,

Kamran

#DecisionOptimization
#OPLusingCPLEXOptimizer

Hi,

in your model I see

if (feasblity[i] = 0)

which is probably wrong and I think you meant

if (feasblity[i] == 0)

regards

#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: Kamran_Sr

Hi Alex,

Many Thanks for your reply. I have modified the if statement as you advised. However, this issue is not resolved yet.

I have attached the model files. I really appreciated any further advice.

 

Regards.

 

#DecisionOptimization
#OPLusingCPLEXOptimizer

Hi,

just to be sure have you fixed the 2 errors:

else if (feasblity[i] == 0)
           { writeln("break"); break; }        
        }
          if (feasblity[i] == 0)       
           { writeln("break"); break; }
         }      

instead of

else if (feasblity[i] = 0)
           { writeln("break"); break; }        
        }
          if (feasblity[i] = 0)       
           { writeln("break"); break; }
         }      

?

 

regards

#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: Kamran_Sr

Hi Alex,

 

Yes, I did and the loop was ended. However, the while loop stopped in the iteration in which the value of feasblity[I] was [2 0 0]. But what I need is the loop should be aborted when all elements of feasblity[I] be 0.

Then I tried to use the decision variable directly in the if condition as follows:

else if (model.SlackCap[i] == 0)
{ writeln("break"); break; }
}
if (model.SlackCap[i] == 0)
{ writeln("break"); break; }
}

 

But again the loop stops as this variable return the 0 value for one index(element of array). And as I mentioned I need the loop stops when all elements of this variable be zero, i.e. [0 0 0 0] which will be obtained in the next iteration. So I am not sure how to set the condition in the if function for all elements of array. I would appreciate  any advice.

 

Thanks again for your assistance.

 

Regards 

#DecisionOptimization
#OPLusingCPLEXOptimizer