Why do I get the same numbers in "randn" function?

Every time I open MATLAB the first time I get the same values when I use randn function. This happens only after I first open MATLAB and not when the program is already open and I keep coming back for calculations.

[Todd Flanagan]

Itamar says, "Thanks everyone for the answers. I learned something today"

Answer 1

[Paulo Silva]

You get the same numbers because there's a default seed (0) for the random number generator but you can change it like this:

RandStream.setDefaultStream(RandStream('mt19937ar','seed',sum(100*clock)))

If you want to see what changed do RandStream.getDefaultStream before and after changing the seed.

[Paulo Silva]

When I started this answer there were no other answers, sorry for repeating what was already said.

[Todd Flanagan]

Don't worry about repeats. Just keep the good answers coming.

[Paulo Silva]

:)

[afef]

i tried this :

 seed = sum(100*clock);
 RandStream.setDefaultStream(RandStream('mt19937ar','seed',seed));
 import java.util.Random;
 import java.lang.System;
 gen = Random(seed);

but i got error message "The class RandStream has no Constant property or Static method named 'setDefaultStream' " how to fix it?

[Steven Lord]

We started announcing that you should replace calls to the setDefaultStream method for the RandStream class with calls to setGlobalStream in release R2011a, started warning users in release R2012a, and started having setDefaultStream throw an error in release R2013a.

Answer 2

[Todd Flanagan]

This is by design:

The sequence of numbers produced by randn is determined by the internal state of the uniform pseudorandom number generator that underlies rand, randi, and randn. randn uses one or more uniform values from that default stream to generate each normal value. Control the default stream using its properties and methods. See @RandStream for details about the default stream.

Resetting the default stream to the same fixed state allows computations to be repeated. Setting the stream to different states leads to unique computations, however, it does not improve any statistical properties. Since the random number generator is initialized to the same state every time MATLAB software starts up, rand, randn, and randi will generate the same sequence of numbers in each session until the state is changed.

On that page there is an example of how to change the default stream of numbers using clock:

RandStream.setDefaultStream ... 
     (RandStream('mt19937ar','seed',sum(100*clock)));
randn(1,5)

[Cris Luengo]

Todd beat me to it! :)

Answer 3

[Cris Luengo]

All you need to do is read the documentation for the function randn: http://www.mathworks.com/help/techdoc/ref/randn.html

>> "Since the random number generator is initialized to the same state every time MATLAB software starts up, rand, randn, and randi will generate the same sequence of numbers in each session until the state is changed."

In the examples section on that page you see:

>> "Replace the default stream at MATLAB startup, using a stream whose seed is based on clock, so that randn will return different values in different MATLAB sessions. It is usually not desirable to do this more than once per MATLAB session."

And then the following example code:

RandStream.setDefaultStream ... 
(RandStream('mt19937ar','seed',sum(100*clock)));

Cheers, Cris.