Re: R: random generation of constrained portfolio a llocation weights

[Ferdinando M. Ametrano-3]

Hi Enrico,

>You can do also in this way:
>- 1- generate n random numbers between [0-1]
>- 2- you transform them into constrained number [20%-40%]
>(simply forcing them to enter into that range:
>ex: you get 0.8, it can be seen as
>0.8/1=x/0.2 new= 0.2+x.
>- 3- then you compute all the sum of these numbers.
>And you rescale all the numbers in order to sum to one

after rescaling in step 3 you have no guarantees that the constraints are
still observed.

ciao -- Nando