random cover of a range

bearophile bearophileHUGS at lycos.com
Sat Feb 14 04:50:53 PST 2009


Bill Baxter:
> if you have a load factor (fraction full) of
> f, then it's expected it will take you 1/(1-f) trials to get an empty
> if you just continue picking randomly one after the other.

That goes well with my practical experience. So instead of 90% it may be better to use 85%, this percentage has to be tuned.

Bye,
bearophile



More information about the Digitalmars-d mailing list