home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: sci.math
- Path: sparky!uunet!zaphod.mps.ohio-state.edu!sdd.hp.com!spool.mu.edu!agate!linus!linus.mitre.org!gauss!bs
- From: bs@gauss.mitre.org (Robert D. Silverman)
- Subject: Re: 1+1/2+1/3+1/4+...+1/n
- Message-ID: <1992Nov21.125910.24073@linus.mitre.org>
- Sender: news@linus.mitre.org (News Service)
- Nntp-Posting-Host: gauss.mitre.org
- Organization: Research Computer Facility, MITRE Corporation, Bedford, MA
- References: <92324.132329K3032E2@ALIJKU11.BITNET> <HANCHE.92Nov19123151@ptolemy.ams.sunysb.edu> <BxzMuB.6Gr@cantua.canterbury.ac.nz>
- Date: Sat, 21 Nov 1992 12:59:10 GMT
- Lines: 43
-
- In article <BxzMuB.6Gr@cantua.canterbury.ac.nz> jsv@math.canterbury.ac.nz (Julian Visch) writes:
- :In article <HANCHE.92Nov19123151@ptolemy.ams.sunysb.edu>, hanche@ams.sunysb.edu (Harald Hanche-Olsen) writes:
- :|> >>>>> On Thursday, 19 Nov 1992 13:23:29 CET, Mutter Christoph Johannes
- :|> >>>>> <K3032E2@ALIJKU11.BITNET> said:
- :|>
- :|> Chris> Hallo.
- :|> Chris> I've a problem. I have to calculate the sum 1+1/2+1/3+1/4+...+1/n.
- :|> Chris> The result should be 100. But my computer is far too slow, to solve this
- :|> Chris> problem. I need the index n when the sum > 100.0
- :|> Chris> And that exactly.
- :|> Chris> Has anyone calculated this? (Perhaps on a CRAY|) ?
- :|>
- :|> You won't find your answer by brute force on any present-day computer.
- :|> Consider the fact that this sum is fairly close to log(n)+gamma for
- :|> large n, where gamma is Euler's constant; it follows that the desired
- :|> n is rather close to exp(100-gamma), which has the approximate value
- :|> 15092688622113788323693563264538101449859497.364099... (according to
- :|> Maple). Some finesse (asymptotic analysis) is clearly called for to
- :|> get the exact answer. Hint: Estimate the area between the step
- :|> function 1/[t] (where [t] is the largest integer <= t) and 1/t for t
- :|> between n and infinity.
- :|>
- :|> - Harald
- :
- :Using Maple with 1000 decimal places I found that for
- :n = 15092688622113788323693563264538101449859496 the sum was just below 100
- :
- :i.e
- :> evalf(sum(1./n,n=1..15092688622113788323693563264538101449859496));
- :99.999999999999999999999999999999999999999999942747074071711130782941148738689\
- :110197701561042496005255810450666769067029302578130750912902932624408446045107\
-
- stuff deleted.....
-
- Just HOW did Maple produce this sum? How do you know it is right?
- I'm not saying it is wrong, but I am suspicious. Using H_n = log(n) + gamma
- + O(1/n) [even getting many error terms via Bernoulli polynomials won't
- give explicit constants on the errors], is NOT sufficient.
- --
- Bob Silverman
- These are my opinions and not MITRE's.
- Mitre Corporation, Bedford, MA 01730
- "You can lead a horse's ass to knowledge, but you can't make him think"
-