Re: Randomly initializing a N-element vector

[jalex]

David Koelle writes:

> I would suggest that you start generating a number between 0.0 and 1.0, then
> subtract the random value from 1.0 so next time you pick a random number
> between 0.0 and 1.0-random, and so on (essentially, to modify Rule B above:
> "The difference of all elements from 1 is 0")
> 
> x = 1.0
> for (i=0; i<n-1; i++) {
>   statevector[i] = random number between 0.0 and x;
>   x = x - r;
> }
> statevector[n] = x;

I thought of this, but doesn't it tend to bias small values of i over
large values of i?  It seems to me that statevector[0] has a greater
chance of being close to 1.0 than statevector[n], since it only takes
a single high "roll of the dice" to set statevector[0] near 1.0, but
it takes n-1 low rolls to get statevector[n] near 1.0

Jason


                  ==================================
   Swarm-Modelling is for discussion of Simulation and Modelling techniques
   esp. using Swarm.  For list administration needs (esp. [un]subscribing),
   please send a message to <address@hidden> with "help" in the
   body of the message.
                  ==================================