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”).

Length of the longest monochromatic arithmetic progressions of difference n in the Rudin-Shapiro sequence (A020985).
1

%I #38 Jun 30 2024 03:14:34

%S 4,4,5,4,6,5,9,4,9,6,15,5,6,9,10,4,10,9,12,6,10,15,13,5,12,6,12,9,12,

%T 10,19,4,18,10,13,9,15,12,22,6,12,10,15,15,12,13,9,5,12,12,15,6,13,12,

%U 13,9,10,12,9,10,18,19,33,4,34,18,10,10,10,13,12,9

%N Length of the longest monochromatic arithmetic progressions of difference n in the Rudin-Shapiro sequence (A020985).

%C Also applies to the other versions of Rudin-Shapiro sequence (e.g., A020987).

%C For n < 2^k the inequality a(n) <= 2^(k+1) holds, and a monochromatic arithmetic progression of length a(n) and difference n appears within 10*4^k initial terms of the Rudin-Shapiro sequence (A020985). More generally, if a(n) <= 2^m, then such a progression appears within 5*2^(k+m) initial terms. Conversely, if the maximal length of a progression within 5*2^(k+m) initial terms is <= 2^m, then also a(n) <= 2^m. These properties follow from the referenced paper by Sobolewski. - _Bartosz Sobolewski_, Jun 17 2024

%H Bartosz Sobolewski, <a href="/A364995/b364995.txt">Table of n, a(n) for n = 1..10000</a>

%H Ibai Aedo, Uwe Grimm, Yasushi Nagai, and Petra Staynova, <a href="https://doi.org/10.1016/j.tcs.2022.08.013">Monochromatic arithmetic progressions in binary Thue-Morse-like words</a>, Theor. Comput. Sci., 934 (2022), 65-80; preprint: <a href="https://arxiv.org/abs/2101.02056">On long arithmetic progressions in binary Morse-like words</a>, arXiv:2101.02056 [math.CO], 2021.

%H Bartosz Sobolewski, <a href="https://arxiv.org/abs/2204.05287">On monochromatic arithmetic progressions in binary words associated with pattern sequences</a>, arXiv:2204.05287 [math.CO], 2023.

%e For n = 3, let r(i) be the i-th term of the Rudin-Shapiro sequence (A020985). We have r(28) = r(31) = r(34) = r(37) = r(40), and no k and m > 5 exist such that r(k) = r(k+3) = r(k+2*3) = ... = r(k+(m-1)*3). So a(3)=5.

%t a[n_] := a[n] = If[EvenQ[n], a[n/2], Max[Map[Length, Split /@ Table[RudinShapiro[m n + j], {j, 1, n}, {m, 0, 10*4^(Floor[Log2[n]] + 1)/n}], {2}]]];

%t Table[a[n], {n, 1, 72}] (* _Bartosz Sobolewski_, Jun 17 2024 *)

%Y Cf. A020985, A020987.

%Y Cf. A342818 (analog for the Thue-Morse sequence).

%K nonn

%O 1,1

%A _Gandhar Joshi_, Aug 15 2023

%E a(33)-a(34) from Sobolewski added by _Gandhar Joshi_, Apr 30 2024

%E Corrected and extended by _Bartosz Sobolewski_, Jun 17 2024