mersenneforum.org Help needed for a sequence
 Register FAQ Search Today's Posts Mark Forums Read

 2014-01-03, 07:16 #1 MichelMarcus   Nov 2011 Saint Maur, France 2×52 Posts Help needed for a sequence I would like to create the sequence of numbers whose aliquot sequence is always increasing until it reaches a prime number. Among the terms one would find for instance 324 (324->523) and 392 (392->463). But also 9112 (9112->9248->10093) or 10712 (10712->11128->11552->12451). Another term would be 14952. What would be the best way to proceed? Many thanks. Michel
 2014-01-03, 10:09 #2 fivemack (loop (#_fork))     Feb 2006 Cambridge, England 6,361 Posts It's a nice question, and leads to some pretty tree-graphs. But I'm not quite sure what you mean by 'the best way to proceed' ... it would be nice to back-track but I don't know a clever way to do that. Code: for(a=5,100000000, if(!(isprime(a)), \ ct=0; aa=a; happy=1; while(happy==1, \ ct=1+ct; b=sigma(aa)-aa; if(b<=aa,happy=0,if(isprime(b),happy=2)); if(ct>50,happy=-1); aa=b); \ if(happy==2,print([a,aa,happy,ct])))) in pari/gp is a stupidly simplistic approach; there are aliquot sequences which keep growing as far as anyone has followed them, so you have to stick in the arbitrary limit somewhere, and if you start finding completions at ct=50 (I've not seen ct>23) then increase the value of 50 and try again.
2014-01-03, 10:30   #3
garambois

Oct 2011

22×5×17 Posts

Quote:
 Originally Posted by MichelMarcus I would like to create the sequence of numbers whose aliquot sequence is always increasing until it reaches a prime number. Among the terms one would find for instance 324 (324->523) and 392 (392->463). But also 9112 (9112->9248->10093) or 10712 (10712->11128->11552->12451). Another term would be 14952. What would be the best way to proceed? Many thanks. Michel
Hello Michel,

I think your idea is very interesting !
I don't know what's the best way to proceed !
All the terms of your sequences will be even numbers (because increasing) except the last.
The last term is prime, so it is odd.
I'm sur that the last but one term of your sequences is of the form m^2 or 2*m^2.
See the proof here (sorry in french) : http://www.aliquotes.com/changement_parite.pdf

Jean-Luc

 2014-01-03, 12:01 #4 MichelMarcus   Nov 2011 Saint Maur, France 1100102 Posts Yes, I ran into these kind of problems. With an arbitrary limit, the first hurdles I got were 966 and 1134. With factordb, I saw that 966 apparently grows monotonously until term 696 that is deficient. Also, the graph of 1134 shows that it is going up and down. But the process is quite lengthy and not very easy to do. Would it be possible to automate it? You wrote that there are aliquot sequences which keep growing as far as anyone has followed them. Do you know the smallest starting term of these sequences? Thanks. Michel
2014-01-03, 12:34   #5
garambois

Oct 2011

34010 Posts

Quote:
 Originally Posted by MichelMarcus Yes, I ran into these kind of problems. With an arbitrary limit, the first hurdles I got were 966 and 1134. With factordb, I saw that 966 apparently grows monotonously until term 696 that is deficient. Also, the graph of 1134 shows that it is going up and down. But the process is quite lengthy and not very easy to do. Would it be possible to automate it? You wrote that there are aliquot sequences which keep growing as far as anyone has followed them. Do you know the smallest starting term of these sequences? Thanks. Michel
I know several aliquot sequences which keep growing as far as anyone has followed them, for exemple : 5352 or 9336. I calculate those two sequences which are reserved for me on this forum on "Aliquot sequence reservations".
Or another exemple with 19560 (which is reserved for Paul Zimmermann) and which is growing at each iteration with a factor at least 2.
Jean-Luc

 2014-01-03, 13:22 #6 MichelMarcus   Nov 2011 Saint Maur, France 1100102 Posts These 3 sequences, that you're telling me about, they never encounter a deficient number? Is that so ?
2014-01-03, 13:48   #7
garambois

Oct 2011

5248 Posts

Quote:
 Originally Posted by MichelMarcus These 3 sequences, that you're telling me about, they never encounter a deficient number? Is that so ?
Yes that's right and I know a lot of others aliquot sequences like that !
But they are all "OE-aliquot sequences" and perhaps if we continue, they will go down ! Nobody can know before calculs will be done !
Jean-Luc

2014-01-03, 14:03   #8
garambois

Oct 2011

5248 Posts

Quote:
 Originally Posted by garambois Hello Michel, I think your idea is very interesting ! I don't know what's the best way to proceed ! But if this can help you : All the terms of your sequences will be even numbers (because increasing) except the last. The last term is prime, so it is odd. I'm sur that the last but one term of your sequences is of the form m^2 or 2*m^2. See the proof here (sorry in french) : http://www.aliquotes.com/changement_parite.pdf Jean-Luc
Sorry, more complicated I knew !
I can just tell you that in a aliquot sequence, if you want the terms of the sequence change their parity, you have to pass by a term which is of the form m^2 or 2*m^2 (m an integer).
So if you begin an aliquot sequence on an even number, it's possible you have a term which is of the forme m^2 or 2*m^2, an next, the terms will be odd. And you can have several odd terms before have a prime to finish the sequence !
If you begin your sequence on a odd term, you can finish your sequence by a prime without have a term of the forme m^2 or 2*m^2 because not necessary to change the parity of the terms.

I hope I am clear because my bad english !

Jean-Luc

 2014-01-03, 14:54 #9 science_man_88     "Forget I exist" Jul 2009 Dumbassville 26×131 Posts I made a PARI code to go backwards once but garambois is a lot better at these things. if you want I can tell you how I came to my code and you could make your own. Last fiddled with by science_man_88 on 2014-01-03 at 14:54
 2014-01-03, 15:11 #10 MichelMarcus   Nov 2011 Saint Maur, France 628 Posts I wrote something like: prev(n) = {for (i=1, n, if ((sigma(i) - i) == n, return (i));); return (0);} Going back to factordb, when I look at say 966, there is a button to download the corresponding elf file. Is there a place where these elf files can be downloaded from ?
2014-01-03, 15:11   #11
fivemack
(loop (#_fork))

Feb 2006
Cambridge, England

143318 Posts

Quote:
 Originally Posted by MichelMarcus These 3 sequences, that you're telling me about, they never encounter a deficient number? Is that so ?
That's unknown - factordb contains, as far as I'm aware, all known knowledge about those sequences - and I suspect it's false.

I am running 30-terms on [1,1e10] to see if I get any more interesting cluster points; it should be done by Monday. 166681535623 looks interesting (26th iterate of 119785014)

 Similar Threads Thread Thread Starter Forum Replies Last Post sweety439 And now for something completely different 17 2017-06-13 03:49 devarajkandadai Math 1 2007-08-25 15:23 devarajkandadai Math 3 2007-03-20 19:43 roger Puzzles 16 2006-10-18 19:52 biwema Puzzles 13 2004-06-11 02:05

All times are UTC. The time now is 07:06.

Thu Oct 1 07:06:21 UTC 2020 up 21 days, 4:17, 0 users, load averages: 1.76, 1.65, 1.62