|
|
A195166
|
|
Numbers expressible as 2^a - 2^b, with 0 <= b < a, such that n^a - n^b is divisible by 2^a - 2^b for all n.
|
|
0
|
|
|
1, 2, 6, 12, 30, 24, 60, 120, 252, 240, 504, 16380, 32760, 65520
(list;
graph;
refs;
listen;
history;
text;
internal format)
|
|
|
OFFSET
|
1,2
|
|
COMMENTS
|
1 = 2^1 - 2^0. (n^1 - n^0)/1 : A000027
2 = 2^2 - 2^1. (n^2 - n^1)/2 : A000217
6 = 2^3 - 2^1. (n^3 - n^1)/6 : A000292
12 = 2^4 - 2^2. (n^4 - n^2)/12 : A002415
30 = 2^5 - 2^1. (n^5 - n^1)/30 : A033455
24 = 2^5 - 2^3. (n^5 - n^3)/24 : A006414
60 = 2^6 - 2^2. (n^6 - n^2)/60 : A213547
120 = 2^7 - 2^3. (n^7 - n^3)/120 : A114239
252 = 2^8 - 2^2. (n^8 - n^2)/252 :
240 = 2^8 - 2^4. (n^8 - n^4)/240 : A078876
504 = 2^9 - 2^3. (n^9 - n^3)/504 :
16380 = 2^14 - 2^2. (n^14 - n^2)/16380 :
32760 = 2^15 - 2^3. (n^15 - n^3)/32760 :
65520 = 2^16 - 2^4. (n^16 - n^4)/65520 :
"The Mod Set Stanford University" and Carl Pomerance independently noted that the completeness of this sequence follows from a result of Schinzel on primitive prime factors of sequences a^n - b^n in the remark to a problem of Harry Ruderman asking whether if 2^a - 2^b divides 3^a - 3^b, then 2^a - 2^b belongs to this sequence.
The Mod Set verified that if a > b, 2^a - 2^b divides 3^a - 3^b, but 2^a - 2^b does not belong to this sequence, then a - b > 1900. Ram Murty and Kumar Murty proved that there are only finitely many natural numbers a, b such that 2^a - 2^b divides 3^a - 3^b.
Qi Sun and Ming Zhi Zhang also showed that if a > b and n^a - n^b is divisible by 2^a - 2^b for all n, then 2^a - 2^b belongs to this sequence. (End)
|
|
LINKS
|
Harry Ruderman, Problem E2468, Amer. Math. Monthly 81 (1974), p. 405.
|
|
EXAMPLE
|
a(3) = 6 belongs to this sequence since (n^3 - n)/6 = C(n+1, 3) = A000292(n-1).
|
|
CROSSREFS
|
|
|
KEYWORD
|
nonn,fini,full
|
|
AUTHOR
|
|
|
STATUS
|
approved
|
|
|
|