Decision Optimization

Hi,

in order to cope with simulation, robustness, Monte Carlo ... random functions are very useful.

In OPL, we have 4 random generators and with 3 of them you can set the random seed.

Let me list them and then show examples on how to set random seeds and then use them.

• Cplex random seed
• CPO random seed
• OPL random seed
• OPL scripting random

int sr=srand(2); // OPL random seed

execute
 {
 sr;
 writeln("OPL random seed=",sr);
 }
 
int oplrandomnumber=rand(20);

execute
{
writeln("OPL random number =",oplrandomnumber);
}
 
execute
{
cplex.randomseed=3; // CPLEX random seed
}

execute
{
writeln("scripting random function = ",Math.random());  // scripting random function
writeln("cplex random seed =",cplex.randomseed);
}

gives

OPL random seed=2
OPL random number =5
scripting random function = 0.001251221
cplex random seed =3

and

using CP;

int sr=srand(2); // OPL random seed
 
 execute
 {
 sr;
 writeln("OPL random seed=",sr)
 }
 
 int oplrandomnumber=rand(20);

execute
{
writeln("OPL random number =",oplrandomnumber);
}
 
 
execute
{
cp.param.RandomSeed=4; // CPO random seed
}

execute
{
writeln(Math.random());  // scripting random function
writeln("cpo random seed=",cp.param.randomseed);

}

gives

OPL random seed=2
OPL random number =5
0.001251221
cpo random seed=4

regards

A 5th way is to call a function outside OPL that can be in java or python or any language.

execute{ 
writeln("5 random numbers");
var rnd = IloOplCallJava("java.util.Random", "<init>", "()");
rnd.setSeed(1);	
for(var i=1;i<=5;i++)
{			    			  
 var t = rnd.nextDouble();
 writeln(t);
}
writeln("Gaussian");
for(var i=1;i<=5;i++)
{			    			  
 var t = rnd.nextGaussian();
 writeln(t);
}
}

in https://github.com/AlexFleischerParis/oplscripting/blob/main/IloOplCallJava.mod

------------------------------
[Alex] [Fleischer]
[Data and AI Technical Sales]
[IBM]
------------------------------