Originally posted by: CMP-SFL

Dear all,

In a linear programming problem, I want to use the root square and logarithm functions.

I found a list of all functions in the following page:

http://pic.dhe.ibm.com/ infocenter/cosinfoc/v12r2/ index.jsp?topic=%2Filog.odms. ide.help%2FContent% 2FOptimization% 2FDocumentation%2FOPL_Studio% 2F_pubskel%2Fglobals% 2Feclipse_and_xplatform%2Fps_ opl749.html

Two functions "log" and "sqrt" fulfill the needs. However, I don't know how to use them. Could you please help me on that?

Below, I'have put the reduced models to illustrate the case.

// Begin of Model 1
minimize sum(m in Unit) varaiable2[m] ;

subject to
{
forall(m in Unit) varaiable2[m] >= (varaiable1[m] - Parameter[m])^(1/2) ;
}

main {
    cplex.solve();
    var UB = cplex.getObjValue();
    var LB = cplex.getBestObjValue();
    }
// End of Model 1




// Begin of Model 2
minimize sum(m in Unit) varaiable2[m] ;

subject to
{
forall(m in Unit) log(varaiable2[m]) >= log (varaiable1[m] - Parameter[m]) ;
}

main {
    cplex.solve();
    var UB = cplex.getObjValue();
    var LB = cplex.getBestObjValue();
    }
// End of Model 2

Many thanks in advance,

Best regards

Mehdi

#DecisionOptimization
#OPLusingCPLEXOptimizer

Hi,

here an example of the syntax:

float s=sqrt(100);
float l=log(100);

execute
{
writeln(s);
writeln(l);
}

 

 

If you want to use the square root or the log of a decision variable, maybe you should rather use CPO and write use CP; at the beginning of your model

regards

 

#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: CMP-SFL

Hi,

Thanks for the answer.

 

I have never used CP optimizer. By looking at OPL examples, I added using CP; at the beginning of the model, and I got other errors. I think, the syntax must be changed!

The models I use are simple. I put again the gist of the models below. Do you know how to put the model in the right format in order to use sqrt and log functions? Thanks again!

// Begin of Model 1

float Parameter [Unit]=...;

dvar float+ varaiable1[Unit];

dvar float+ varaiable2[Unit];

 

 

minimize sum(m in Unit) varaiable2[m] ;

 

subject to

{

forall(m in Unit) varaiable2[m] >= (varaiable1[m] - Parameter[m])^(1/2) ;

}

 

main {

    cplex.solve();

    var UB = cplex.getObjValue();

    var LB = cplex.getBestObjValue();

    }

// End of Model 1

 

// Begin of Model 2

float Parameter [Unit]=...;

 

dvar float+ varaiable1[Unit];

dvar float+ varaiable2[Unit];

 

minimize sum(m in Unit) varaiable2[m] ;

 

subject to

{

forall(m in Unit) log(varaiable2[m]) >= log (varaiable1[m] - Parameter[m]) ;

}

 

main {

    cplex.solve();

    var UB = cplex.getObjValue();

    var LB = cplex.getBestObjValue();

    }

// End of Model 2

#DecisionOptimization
#OPLusingCPLEXOptimizer