Originally posted by: pareto

Dear All,

I intend to read warm-start values for 8 different decision variables. I do that typically as the way presented in the example. But, seemingly, I overwrite the previous vectors each time. Thats what I have tired:

 

main{ 

...

var data= new IloOplDataSource(dataFileName);

opl.addDataSource(data);

opl.generate();

var vectors = newIloOplCplexVectors();

vectors.attach(opl.B, opl.position);

vectors.setVectors(Cplex);

var vectors2 = newIloOplCplexVectors();

vectors2.attach(opl.k, opl.value);

vectors2.setVectors(Cplex);

... }

I have also tried using only one vectors.setVectors(Cplex) in the end, but it still just reads warm-start for only one decision variable family.

Thanks for help

 

#DecisionOptimization
#OPLusingCPLEXOptimizer

Hi

have you had a look at the example warmstart.mod in the OPL examples ?

regards

#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: pareto

Hello Alex,

I have looked at it. But there, the warm-start just reads one decision variable (x) family, I also would like to read a warm-start for y variables.

If I do that like this:

var vectors = newIloOplCplexVectors();

vectors.attach(opl.B, opl.position);

vectors.setVectors(Cplex);

var vectors2 = newIloOplCplexVectors();

vectors2.attach(opl.k, opl.value);

vectors2.setVectors(Cplex);

...

It just considers one warm-start values (I guess it overwrites the vector) or sth else. I want to give initial values for both x and y DV family.

 

Regards,

 

#DecisionOptimization
#OPLusingCPLEXOptimizer

Hi,

I changed warmstart.mod in order to take into consideration 2 decision variables:

range r = 1..10;
dvar int+ x[r];
dvar int+ y[r];

dvar int+ x2[r];
dvar int+ y2[r];
// The following array of values (defined as data) will be used as
// a starting solution to warm-start the CPLEX search.
float values[i in r] = (i==5)? 10 : 0;
float values2[i in r] = (i==4)? 10 : 0;


minimize
  sum( i in r ) x[i] + sum( j in r ) y[j]+sum( i in r ) x2[i] + sum( j in r ) y2[j];
subject to{
  ctSum:    
    sum( i in r ) x[i] >= 10;
  forall( j in r )
    ctEqual:
      y[j] == j;
      
      ctSum2:    
    sum( i in r ) x2[i] >= 10;
  forall( j in r )
    ctEqual2:
      y2[j] == j;
}

main{
  thisOplModel.generate();  
  var def = thisOplModel.modelDefinition;   
  // Default behaviour
  writeln("Default Behaviour");
  var cplex1 = new IloCplex();
  var opl1 = new IloOplModel(def, cplex1);
  opl1.generate();
  cplex1.solve();   
  writeln(opl1.printSolution());
  // Setting initial solution
  writeln("Setting initial solution");
  var cplex2 = new IloCplex();
  var opl2 = new IloOplModel(def, cplex2);
  opl2.generate();
  var vectors = new IloOplCplexVectors();
  // We attach the values (defined as data) as starting solution
  // for the variables x.
  vectors.attach(opl2.x,opl2.values);
  vectors.attach(opl2.x2,opl2.values2);
  vectors.setVectors(cplex2);   
  cplex2.solve();   
  writeln(opl2.printSolution());

  opl1.end();
  cplex1.end();
  opl2.end();
  cplex2.end();
  0;
}

 

regards

#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: pareto

Hello Alex,

I didnt understand why you have duplicated the decision variables, let me give you a more precise example with a similar model:

if I write the code like this: 

MIP start 'm1' defined initial solution with objective 36.0000. (which doesnt take values2(y) into account) . It is just objective value(x) when values are considered.

If I change the order in main and write:

 vectors.attach(opl2.y,opl2.values2);

vectors.attach(opl2.x,opl2.values);

Then;

MIP start 'm1' defined initial solution with objective 14.0000. (now, it doesnt consider values(x warm-start variables), but y variables ). So, the warm-start is changed.

Normally, if I give warm-start for both x and y. I should get a MIP start of 40.000

Maybe I didnt get it completely, but I didnt understand it

 

Kind Regards

 

#DecisionOptimization
#OPLusingCPLEXOptimizer

Hi,

in both cases I get get a MIP start of 40.000

Which version are you using ? Cplex 12.6 ?

regards

#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: pareto

Hello again,

I use 12.5.1

#DecisionOptimization
#OPLusingCPLEXOptimizer

Indeed I have your issue with 12.5.1

Can you move to 12.6 then ?

regards

#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: Preemptive

 

Hi, 

F[i][v][t] denotes arrival time  at node i by vehicle v after traversing t-th arc. So, if x[i][j][v][t]==1 than F[j][v][t] must be equal to F[i][v][t-1] + t[i][j] since vehicle v arrives at node i after traversing (t-1) arc. so i have this constraint : 

 

 

 

#DecisionOptimization
#OPLusingCPLEXOptimizer