A103314: Conjectures for
David Wilson
davidwwilson at comcast.net
Sun May 8 09:21:50 CEST 2005
Let a = A103314. I conjecture that
[1] a(n) = a(s(n))^(a(n)/s(n))
where s(n) = A007947(n) is the largest squarefree number dividing n. If I
knew that all zero-sum subsets of roots of 1 were gotten by set operations
on cyclical zero-sum subsets, I could probably cob together a proof of [1].
Since s(n) is squarefree, [1] reduces the problem of computing a(n) to that
of computing a(n) for the squarefree integers. The squarefree numbers are
products of distinct primes, and I suspect that expressions for a(n) for
squarefree n can be formulated based on the number of primes in the product.
For 0, 1, and 2 primes, we already have such expressions, namely:
[2] a(1) = 1
[3] a(p) = 2
[4] a(pq) = a(pq) = 2^p+2^q-2
(here and henceforth p and q stand for primes).
[2] is of limited application. [1] and [3] together give
[5] a(p^k) = 2^p^(k-1) (k >= 1)
while [1] and [4] give
[6] a(p^j q^k) = (2^p+2^q-2)^(p^(j-1) q^(k-1)) (j, k >= 1).
[5] and [6] allow us to compute A103314(n) for any n with up to 2 prime
divisors.
The remaining identies noted on A103314(n) also follow from [1]-[4]. These
are
[7] a(2^n) = 2^2^(n-1)
This is a special case of [5] with p = 2.
[8] a(2n) = a(n)^2 (n even)
More succinctly, a(4n) = a(2n)^2. To show this, let 2n = (2^k)m (k >= 1, m
odd), giving 4m = (2^(k+1))m. Note that s is multiplicative, so s(2n) =
s((2^k)m) = s(2^k)s(m) = 2s(m), and similarly s(4m) = 2s(m). Therefore
a(2n) = a(s(2n))^(2n/s(2n)) = a(2s(m))^(2n/(2s(m))) = a(2s(m))^(n/s(m))
a(4n) = a(s(2n))^(4n/s(4n)) = a(2s(m))^(4n/(2s(m))) = a(2s(m))^(2n/s(m))
from which a(4n) = a(2n)^2 is evident.
[9] a(2p) = 2^p+2
follows immediately from [4] as noted.
Also, [6] gives a(2^j 3^k) = 10^(2^(j-1) 3^(k-1)) (j, k >= 1), accounting
for the powers of 10 we see A103314. It predicts that a(n) = 10^(n/6) when
n is of the form 2^j 3^k (j, k >= 1).
- David W. Wilson
"Truth is just truth -- You can't have opinions about the truth."
- Peter Schickele, from P.D.Q. Bach's oratorio "The Seasonings"
More information about the SeqFan
mailing list