Low-Discrepancy Sequences with Accept-Reject Schemes


helix
06-20-2007, 01:28 PM
When transforming uniform variates to normal variates (for example), is it legitimate to utilise accept-reject type algorithms such as Box-Muller?

Given that I will be discarding some numbers in the sequence, will this destroy the low-discrepancy properties of the sequence?

More importantly, will I still get the same answer out of my MonteCarlo?

Many thanks!!