View Single Post
Old 2014-12-05, 17:32   #1
XYYXF
 
XYYXF's Avatar
 
Jan 2005
Minsk, Belarus

24×52 Posts
Default Consecutive p-smooth integers

Let's define L(n, k) as the largest prime factor of product

n*...*(n+k)

of k+1 consecutive integers, starting at positive integer n.

So we have, for example,
L(4374, 1) = 7
L(48, 2) = 7
L(350, 2) = 13
L(138982582998, 2) = 103
L(61011223, 3) = 163
L(23931257472314, 3) = 631
L(1517, 4) = 41
L(3294850, 5) = 239
L(1913253200, 8) = 3499
L(8559986129664, 12) = 58393
L(48503, 14) = 379

Conjecture:
as n goes to infinity,

lim inf L(n, k) / (log n)^2 = C_k

The very rough estimates of constants C_k are:
C_1 ~ 0.0376
C_2 ~ 0.258
C_3 ~ 0.907
C_4 ~ 2.46
C_5 ~ 5.16
C_6 ~ 11.4
C_7 ~ 19
C_8 ~ 42
C_9 ~ 70
C_10 ~ 140
C_11 ~ 200
C_12 ~ 250
C_13 ~ 380
C_14 ~ 430
C_15 ~ 460

Some successive minima of L(n, k) are shown there:
http://oeis.org/A193943
http://oeis.org/A193944
http://oeis.org/A193945
http://oeis.org/A193946
http://oeis.org/A193947
http://oeis.org/A193948
http://oeis.org/A199407
http://oeis.org/A200566
http://oeis.org/A200567
http://oeis.org/A200568
http://oeis.org/A200569
http://oeis.org/A200570
http://oeis.org/A209837
http://oeis.org/A209838
http://oeis.org/A209839

Any suggestions on the conjecture? Does it depend on other
known ones like Twin prime conjecture or ABC conjecture?

Great thanks for any information on the subject.
XYYXF is offline   Reply With Quote