home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!pipex!bnr.co.uk!uknet!mcsun!julienas!irisa!saouter
- From: saouter@irisa.fr (Saouter Yannick)
- Newsgroups: sci.math
- Subject: Square roots and continued fractions
- Message-ID: <1992Dec21.110741.17662@irisa.fr>
- Date: 21 Dec 92 11:07:41 GMT
- Sender: news@irisa.fr
- Organization: IRISA, Rennes (Fr)
- Lines: 23
-
-
-
- I consider the following algorithm to evaluate the square root of M:
-
- x_0=1
-
- x_{n+1}=\frac{x_n}{2}+\frac{M}{2 . x_n}
-
- coming from Newton method.
-
- In my previous courses one of my math teacher says that this algorithm is very fast and is
- even a sub-series of the series of continued fraction for M integers.
-
- It seems to be true for M=2 and 3 and is likely to be easily demonstrable. But for higher
- number, this basic algorithm seems not to behave in this way.
-
- If we initially take x_0=int(\sqrt{x}) where int denotes the integer part, it seems again
- to give a sub-series of continued fraction.
-
-
- Is there any pointers or bibliography to this kind of problems ?
-
- Yannick Saouter - saouter@geocub.greco-prog.fr
-