login

Year-end appeal: Please make a donation to the OEIS Foundation to support ongoing development and maintenance of the OEIS. We are now in our 61st year, we have over 378,000 sequences, and we’ve reached 11,000 citations (which often say “discovered thanks to the OEIS”).

A359028
Integers m such that A006218(m+1)/(m+1) > A006218(m)/m.
1
1, 2, 3, 5, 7, 8, 9, 11, 13, 14, 15, 17, 19, 20, 21, 23, 25, 26, 27, 29, 31, 32, 33, 34, 35, 37, 38, 39, 41, 43, 44, 47, 49, 51, 53, 55, 59, 62, 63, 65, 67, 69, 71, 74, 75, 77, 79, 80, 83, 87, 89, 91, 95, 97, 98, 99, 101, 103, 104, 107, 109, 111, 113, 115, 116, 119, 123, 125, 127, 129
OFFSET
1,2
COMMENTS
Equivalently: Indices m such that f(m + 1) > f(m) where f(m) = Sum_{k=1..m} d(k) / m, where d(k) is the number of divisors of k (A000005).
This sequence comes from a problem proposed by South Africa during the 47th International Mathematical Olympiad, in 2006 at Ljubljana, Slovenia, but not used for the competition (see link).
In fact, the problem asked for a proof that, for the sequence {f(m)} defined by f(m) = (1/m) * ([m/1] + [m/2] + ... + [m/m]), where [x] denotes the integer part of x,
(a) f(m + 1) > f(m) occurs infinitely often (these are the terms m of this sequence),
(b) f(m + 1) < f(m) occurs infinitely often (see A359029).
Differs from A047255 when a(24) = 34 while A047255(24) = 35.
Some results:
1. For every m, f(m) = (1/m) * ([m/1] + [m/2] + ... + [m/m]) proposed in the problem is the arithmetic mean of d(1), d(2), ..., d(m) = A006218(m)/m.
2. f(m + 1) > f(m) is equivalent to d(m + 1) > f(m).
3. Each m = c - 1, where c is a highly composite number (A002182) is a term.
Proof: in this case, d(m+1) = d(c) > max{d(1), ..., d(m)}; as f(m) = (d(1)+...+d(m)) / m < m*d(c)/m = d(c), it follows that d(m+1) = d(c) > f(m).
4. As there are infinitely many highly composite numbers, that also proves that f(m + 1) > f(m) occurs infinitely often, answer to IMO problem (a).
5. There exist other terms not of the form A002182 - 1: 2, 7, 8, 9, 13, 14, 15, ...
Note that f(m) = f(m+1) is possible iff f(m) = tau(m+1), so f(m) must be an integer (A050226) but this is not sufficient. The only known term such that f(m) = f(m+1) is at m=4, with f(4) = 2 and f(5) = tau(5) = 2.
LINKS
47th International Mathematical Olympiad, Slovenia, 2006, Problem N3, page 57, Shortlisted problems with solutions.
EXAMPLE
f(7) = (d(1)+d(2)+d(3)+d(4)+d(5)+d(6)+d(7)) / 7 = (1+2+2+3+2+4+2) / 7 = 16/7 < f(6) = (d(1)+d(2)+d(3)+d(4)+d(5)+d(6)) / 6 = (1+2+2+3+2+4) / 6 = 14/6 = 7/3, so 6 is a term.
MAPLE
with(numtheory):
for n from 1 to 100 do
m := (1/(n+1))*sum(tau(k), k=1..n+1) - (1/n)*sum(tau(k), k=1..n);
if m>0 then print(n); else fi; od:
MATHEMATICA
With[{m = 130}, Position[Differences[Accumulate[DivisorSigma[0, Range[m]]]/Range[m]], _?(# > 0 &)] // Flatten] (* Amiram Eldar, Dec 12 2022 *)
CROSSREFS
KEYWORD
nonn
AUTHOR
Bernard Schott, Dec 12 2022
STATUS
approved