home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: sci.math
- Path: sparky!uunet!utcsri!skule.ecf!torn!news.ccs.queensu.ca!mast.queensu.ca!dmurdoch
- From: dmurdoch@mast.queensu.ca (Duncan Murdoch)
- Subject: Diophantine approximation
- Message-ID: <dmurdoch.30.713816259@mast.queensu.ca>
- Lines: 25
- Sender: news@knot.ccs.queensu.ca (Netnews control)
- Organization: Queen's University
- Date: Fri, 14 Aug 1992 18:17:39 GMT
-
- I'm looking for a pair of reals (x_1,x_2), such that rational approximations
- to the pair of the form (q_1, q_2) = (m_1/n, m_2/n) do as poorly as
- possible, in the sense that the distance from the pair of reals to the
- pair of rationals shrinks at the slowest possible rate as n is allowed to
- increase.
-
- I've looked in Gruber and Lekkerkerker, Geometry of Numbers, and this looks
- like their example I in section 45.1 and example II in section 45.4; in the
- latter, they show that for any choice of (x_1, x_2) there are infinitely
- many pairs (q_1, q_2) as above such that
-
- (x_1 - q_1)^2 + (x_2 - q_2)^2 <= 1/(gamma n^3) (*)
-
- where gamma = sqrt(23)/2, but the assertion is not true for larger gamma.
- In this context, what I think I'm after is a particular pair
- (x_1, x_2) such that there is at best equality in (*), never strict
- inequality; failing that, one where the ratio of the LHS to the RHS is
- asymptotically bounded below by 1.
-
- The reason I have some hope to achieve this is that the corresponding one
- dimensional problem does have solutions, for example x = (sqrt(5) - 1)/2.
-
- The reason I'm asking here is that I can't understand 90% of Gruber and
- Lekkerkerker, so I'm hoping that someone who can read it might know the
- answer.
-