home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!wupost!waikato.ac.nz!canterbury.ac.nz!math!wft
- Newsgroups: sci.math
- Subject: Re: Sums of reciprocals
- Message-ID: <Bzosso.FH9@cantua.canterbury.ac.nz>
- From: wft@math.canterbury.ac.nz (Bill Taylor)
- Date: Wed, 23 Dec 1992 00:46:00 GMT
- References: <ARA.92Dec21193839@camelot.ai.mit.edu>
- Distribution: sci
- Organization: Department of Mathematics, University of Canterbury
- Nntp-Posting-Host: sss330.canterbury.ac.nz
- Lines: 47
-
- In article <ARA.92Dec21193839@camelot.ai.mit.edu>, ara@zurich.ai.mit.edu (Allan Adler) writes:
- |>
- |> Let x be a positive real number.
- |> Let g(n,x) be defined as follows:
- |>
- |> (1) g(n,x) is a strictly increasing sequence of positive integers, possibly
- |> a terminating sequence.
- |> (2) for all n, g(n,x) is the smallest integer consistent with (1)
- |> such that the sum of the reciprocals of g(k,x) for k=1,...,n
- |> does not exceed x.
- |>
- |> I have been told that if x is rational then this leads to a way of writing
- |> x as a finite sum of reciprocals of distinct positive integers. Let the number
- |> of terms in this sum be N=N(x). I have been told that N is also the number of
- |> terms in a minimal representation of x as a sum of reciprocals of positive
- |> integers allowing repetitions.
- |>
- |> Is this true and if so how or where is it proved?
-
- The second statement is false.
-
- Let x = 9/20.
- Then the sequence of g(n,x) is 3, 9, 180.
-
- So N=3, but this is not minimal as x = 1/4 + 1/5 .
-
-
- The first statement is true. That is, for x rational, the process of constructing
- the g(n,x) is certain to terminate. This is easily seen by noting that the
- successive remainders after subtracting the 1/g(n,x) terms are fractions whose
- numerators strictly decrease, so must eventually reach 1.
-
- If x (or its current remainder) is p/q, and is approximated below by 1/n,
-
- i.e. 1/n < p/q < 1/(n-1) , ....(A)
-
- then next remainder = p/q - 1/n
- = (pn-q)/qn and (pn-q) < p from (A). Q.E.D.
- -------------------------------------------------------------------------------
- Bill Taylor wft@math.canterbury.ac.nz
- -------------------------------------------------------------------------------
- All finite math is essentially trivial.
- All continuum math is essentially false.
- -------------------------------------------------------------------------------
-
-
-
-