Originally posted by: FQYA_Majid_Khoshghalb

Hi All

1- how I can update a dvar pos][] after execute, I mean, I want to do some computations in postprocessing and then update pos[][.
how we can update a decision variable after execution?

2- how we can export computations we did in post processing in a .dat file or .txt file?

3- how we can make a matrix in EXECUTE block, because I can make Array, but no information about Matrices.
#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: SystemAdmin

I am not sure if I completely understand your question, but here's some code to get you started. It creates a trivial model, uses the optimal values of the decision variable in the post-processsing (POST_SOLVE) execute block, and then writes some of the values to a txt file in disk. I have also included some code to create a 2-D Array object, but using the same concept you can extend it to more dimensions as your need may be:

range r = 1..5; 

int y[i in r] = i;   dvar int+ x[r]; minimize sum(i in r) x[i]; subject to 
{ forall(i in r) x[i] >= 2*y[i]; 
} execute POST_SOLVE
{ var f = 

new IloOplOutputFile(
"Output.txt"); var modifiedValue=0; f.writeln(
"Testing OUTPUT"); writeln(
"Objective Value = "+cplex.getObjValue()); f.writeln(
"Objective Value = "+cplex.getObjValue()); 

for(var index in r)
{ modifiedValue=2*x[index]; writeln(
"x["+index+
"]="+x[index]+
"; modifiedValue="+modifiedValue); f.writeln(
"x["+index+
"]="+x[index]+
"; modifiedValue="+modifiedValue); 
} 
//extending the Array object in two dimensions to make it a matrix var a2 = 

new Array(); 

for (var i=0; i<3; i++) 
{ a2[i] = 

new Array(); 

for(var j=0; j<3; j++) 
{ a2[i][j] = 10; 
} 
} 

for (var ii=0; ii<3; ii++) 
{ 

for(var jj=0; jj<3; jj++) 
{ writeln(
"a2["+ii+
"]["+jj+
"]="+a2[ii][jj]); f.writeln(
"a2["+ii+
"]["+jj+
"]="+a2[ii][jj]); 
} 
} f.close(); 
}

#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: FQYA_Majid_Khoshghalb

Thanks a lot.
Do you know how we can generate all feasible sequences on every machine in a scheduling problem?
#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: SystemAdmin

> Do you know how we can generate all feasible sequences on every machine
> in a scheduling problem?

You could do so by setting a few CPLEX parameters and then calling the cplex.populate()
method instead of the cplex.solve() method. One such trivial model could look like the
following which prints out the objective value and variable values of each feasible
solution into a file called 'feasibleSolutions.txt':

range r = 1..5; 

int y[i in r] = i; dvar int+ x[r]; minimize sum(i in r) x[i]; subject to 
{ forall(i in r)
{ x[i] >= 2*y[i]; x[i] <= 3*y[i]; 
} 
} main
{ thisOplModel.generate(); 
//set cplex parameters to enumerate all possible feasible solutions cplex.populatelim = 2100000000;
//increase the limit of solutions in the solution pool cplex.solnpoolintensity = 4;
//search for all possible solutions cplex.solnpoolagap=1e+75;
//allow any feasible solution to be part of the solution pool cplex.populate();
//call populate instead of solve var nSols = cplex.solnPoolNsolns; var xRange = thisOplModel.r; var fOut = 

new IloOplOutputFile(
"feasibleSolutions.txt"); 

for(var solIndex=0;solIndex<nSols;solIndex++)
{ thisOplModel.setPoolSolution(solIndex); fOut.writeln(
"Solution#"+(solIndex+1)+
" : Objective="+cplex.getObjValue()); 

for(var xIndex in xRange )
{ fOut.writeln(
"x["+xIndex+
"]="+thisOplModel.x[xIndex]); 
} 
} fOut.close(); 
}

#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: FQYA_Majid_Khoshghalb

Thanks a lot, but my code is using CP;
what can I do for that?
#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: FQYA_Majid_Khoshghalb

In fact, I want to do the same you sent for scheduling problem to have all possible sequence
#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: SystemAdmin

Am sorry about the delay in my response- was on vacation. If your's is a satisfiability problem using CP, you could use the following scheme:

using CP;
range r = 1..5;
int y[i in r] = i;
 
dvar int+ x[r];
subject to {
  forall(i in r){
     x[i] >= 2*y[i];
     x[i] <= 3*y[i];
  }     
} 
main{
    thisOplModel.generate();
    cp.startNewSearch();
    var fOut = new IloOplOutputFile("feasibleSolutions.txt");
    var solIndex=0;
    var xRange = thisOplModel.r;
    while (cp.next()) {
     solIndex++;
     fOut.writeln("Solution#"+solIndex);
     for(var xIndex in xRange ){
        fOut.writeln("x["+xIndex+"]="+thisOplModel.x[xIndex]);
     }                          
    }
    fOut.close();
}

#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: FQYA_Majid_Khoshghalb

Thanks.

Wish the Best.
Majid.
#DecisionOptimization
#OPLusingCPLEXOptimizer