Talk:Las Vegas algorithm

The following is given as the definition of Monte Carlo and Las Vegas by Babai, Cooperman, Finkelstein, Luks and Seress in Fast Monte Carlo Algorithms for Permutation Groups:


 * A Monte Carlo algorithm is a randomized algorithm that may give a wrong answer or report failure

with guaranteed small probability


 * A Las Vegas algorithm is a Monte Carlo algorithm that never gives a wrong answer.

Now this is a bit different than gambling with resource. It seems like ZPP can be defined with either polytime bounded Las Vegas algorithms or with expected polynomial time randomized decision algorithms, giving the same class, but by slightly different machinations. -Ozga 03:23, 5 April 2006 (UTC)

Term Etymology?
I added a reference to the NIST Dictionary of Algorithms and Data Structures definition, although I'm not sure that's the best reference. Does anybody have any idea from where the term originated?

BTW, the NIST definition agrees a little more with "gambling with resources" in that for each run, for the same problem instance, the runtime resource requirements (time/space/etc) will possibly be different. But all programs, etc "gamble with resources", always assuming that there will be enough resources to solve any particular problem instance (and sometimes failing). So I agree, probably ought to consider better wording. Root4(one) 04:58, 12 January 2007 (UTC)

Want more examples
I should like to see more examples of Las Vegas algorithms. For instance, is bogosort a Las Vegas algorithm? Bogosort is very random but is supposed to eventually give the right solution. Right now, there is only one example of a Las Vegas algorithm in the article, and it isn't even particularly well-explained. I'm sure that adding more examples will make this article far more accessible to laymen. Purpy Pupple (talk) 04:43, 6 December 2010 (UTC)
 * Yes, bogosort is a simple example of a Las Vegas algorithm. You might want to create a new section with examples. Please, be WP:BOLD and edit the relevant part(s). Hermel (talk)