home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!munnari.oz.au!comp.vuw.ac.nz!waikato.ac.nz!canterbury.ac.nz!math!wft
- Newsgroups: sci.math
- Subject: Prime constellations.
- Message-ID: <1992Aug27.185156.565@csc.canterbury.ac.nz>
- From: wft@math.canterbury.ac.nz (Bill Taylor)
- Date: 27 Aug 92 18:51:55 +1200
- Distribution: world
- Organization: Department of Mathematics, University of Canterbury
- Nntp-Posting-Host: math.canterbury.ac.nz
- Lines: 70
-
- Some time ago, there was a brief discussion concerning
- "prime constellations", that is , a set of N integers
- ~~~~~~~~~~~~~~~~~~~~ { 0, a2, a3, ..., aN } such that
-
- { p, p+a2, p+a3, ... p+aN } might all be prime, for infinitely many p.
- So that for instance, the TWIN PRIME conjecture concerns constellation { 0, 2 } .
-
- The suggestion was made that constellations might have solutions in
- primes, for either infinitely many primes "p", or only 0 or 1 such "p".
-
- That infinitely-many-primes cases exist, is a famous conjecture,
- connected with Hardy, Bateman and others.
-
- The suggestion that the only finitely-many-solutions could be achieved for
- cases 0 and 1 was soon shot down by many examples,
- e.g. the constellation A = { 0, 2, 8, 14, 26 } has 2 solutions
-
- A+3 = { 3, 5, 11, 17, 29 } and
- A+5 = { 5, 7, 13, 19, 31},
-
- and as A includes all residues [mod 5] there cannot be any further solutions.
-
- Chris Long outlined a nice little ad hoc procedure for generating such
- 2-solution constellations, and asked for the smallest (in some sense)
- constellation that has exactly three solutions.
- ~~~~~~~~~~~~~~~~~~~~~~~
- There seem to have been no further followups. A pity.
-
- To try and resuscitate interest I decided to post my rather boring solution.
-
- I extended Chris Long's method to generate a 3-solution constellation.
- The smallest possible pre-constellation pattern I could find was
-
- 7 13 37 ... OR 7 31 37 ...
- 31 37 61 ... 13 37 43 ...
- 37 43 67 ... 37 61 67 ...
-
- Either of these can be the start of 3 solutions for a constellation of
- cardinality 37, and, provided the 37 entries (on the top line, say)
- include every possible number [mod 37], there can be no further solution.
- Thus we would get a constellation with *exactly* 3 solutions.
-
- As far as I could see, 37 is the smallest possible integer with this property.
- ~~~~~~~~~~~~~~~~~~
- So, I then set Maple to work to find the first 34 extra numbers that would
- give all primes, in all 3 rows, and cover all 37 residues [mod 37].
-
- It found that the left-hand pattern gave the smallest numbers, giving
- a constellation A with
- A+7 =
-
- { 7, 13, 37, 43, 59, 73, 79, 83, 107, 127, 149, 167, 199, 233, 239, 283, 359,
- 409, 419, 479, 547, 563, 617, 797, 829, 953, 1009, 1039, 1063, 1399, 1429, 1669,
- 1723, 1847, 2659, 3583, 3607 }
-
- These, and A+31, and A+37 give all primes, and no other A+p can do so.
-
- And this is (I think) in sense above, the smallest such example.
- ----
- Well; I warned you it would be pretty boring ! There is no problem that
- I can see with extending the method to get 4-fold solutions, but the numbers
- involved are going to be pretty horrific, I would say.
-
- It would be nice to see some further action on this thread !
-
- Cheers,
- --------------------------------------------------------------------------
- Bill Taylor wft@math.canterbury.ac.nz
- --------------------------------------------------------------------------
-
-