|
|
|
|
3, 7, 13, 17, 21, 27, 31, 35, 39, 43, 49, 53, 57, 63, 67, 71, 77, 81, 85, 89, 93, 99, 103, 107, 113, 117, 121, 127, 131, 135, 141, 145, 149, 155, 159, 163, 167, 171, 177, 181, 185, 191, 195, 199, 205, 209, 213, 217, 221, 227, 231, 235, 241, 245, 249, 255
(list;
graph;
refs;
listen;
history;
text;
internal format)
|
|
|
OFFSET
|
1,1
|
|
COMMENTS
|
Conjecture: -1 < n*r - a(n) < 4 for n>=1, where r = (5 + sqrt(17))/2.
Here is a proof of a weak form of the conjecture by Kimberling.
The limit of a(n)/n as n tends to infinity is equal to the inverse 1/f0 of the frequency f0 of the letter 0 in the sequence (a(n)) (see also the argument in A285401).
From the Perron-Frobenius theorem we know that the frequency vector (f0,f1) is the normalized eigenvector associated to the Perron-Frobenius eigenvalue lambda of the incidence matrix M of the generating morphism 0->11, 1->1101.
Here,
M = |0 1|
|2 3|.
The Perron eigenvalue of M is the number lambda = (3+sqrt(17))/2, and [u,v]:=[1,(3+sqrt(17))/2] is an eigenvector with lambda. So
1/f0 = (u+v)/u = (5+sqrt(17))/2,
which identifies the number r in the conjecture.
The fact that (n*r - a(n)) is a bounded sequence can be proved using a general symbolic discrepancy result, Theorem 1 in Adamscewski's 2004 paper.
(End)
|
|
LINKS
|
|
|
EXAMPLE
|
As a word, A285668 = 11011101..., in which 0 is in positions 3,7,13,17,...
|
|
MATHEMATICA
|
s = Nest[Flatten[# /. {0 -> {1, 1}, 1 -> {1, 1, 0, 1}}] &, {0}, 9] (* A285668 *)
Flatten[Position[s, 0]] (* A285669 *)
Flatten[Position[s, 1]] (* A285670 *)
|
|
CROSSREFS
|
|
|
KEYWORD
|
nonn,easy
|
|
AUTHOR
|
|
|
STATUS
|
approved
|
|
|
|