I’ve recently had another result which makes me question the randomness of the numbers being generated by the mersenne twister in MASON.

 

I’ve got two classes that I am setting up the Mersenne twister in separately, because originally I was getting a series of identical numbers in when trying to use the random numbers from the main class.

----------

 

public ABCycle(long seed)        {

           

                        super (new MersenneTwisterFast(seed), new Schedule(5));

            }

 

ABCycle abc = new ABCycle(System.currentTimeMillis());

 

 

-----------------------

 

MersenneTwisterFast random = new MersenneTwisterFast();

 

random.setSeed(System.currentTimeMillis() + id);

                        tp = maxTP * random.nextDouble();

 

------

 

 

Does anything in my code look strange? My results that are emerging have some differentiation between runs, but the standard deviations are miniscule. I feel like one of the randoms is working, but others are not.

 

mike

 

Michael Makowsky

Graduate Student

Department of Economics

George Mason University

703.608.1366

[log in to unmask]

website: http://mason.gmu.edu/~mmakowsk