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”).
%I #23 Jul 28 2023 15:10:40
%S 1,-3,5,-9,18,-36,72,-144,288,-576,1152,-2304,4608,-9216,18432,-36864,
%T 73728,-147456,294912,-589824,1179648,-2359296,4718592,-9437184,
%U 18874368,-37748736,75497472,-150994944,301989888,-603979776,1207959552,-2415919104,4831838208
%N Coefficients in the expansion of 1/([r] + [2r]x + [3r]x^2 + ...); [ ] = floor, r = 3/2.
%C After first 3 terms, agrees with A005010 except for signs; in particular 9 divides a(n) for n >= 3.
%C Suppose r = c/d is a rational number and (a(n)) is the coefficient series for 1/([r] + [2r]x + [3r]x^2 + ...). Let (s(k)) be the increasing sequence of indices n(k) for which a(n(k)) > = 0. In the table below, "yes" indicates that a check of the first 1000 terms indicates that (n(k)) is (eventually) periodic. Column 1 gives selected values of r, and column 2 gives the corresponding coefficient series.
%C 3/2 A279634 yes
%C 4/3 A279675 no
%C 5/3 A279676 no
%C 5/4 A279677 yes
%C 7/4 A279678 yes
%C 6/5 A279778 no
%C 7/5 A279779 no
%C 8/5 A279780 yes
%C 9/5 A279781 no
%H Clark Kimberling, <a href="/A279634/b279634.txt">Table of n, a(n) for n = 0..1000</a>
%H <a href="/index/Rec#order_01">Index entries for linear recurrences with constant coefficients</a>, signature (-2).
%F G.f.: 1/(1 + 3x + 4x^2 + 6x^3 + ...).
%F G.f.: (1 - x) (1 - x^2)/(1 + 2x).
%F E.g.f.: - (1/8) - (3/4)*x + (1/4)*x^2 + (9/8)*exp(-2*x). - _Alejandro J. Becerra Jr._, Feb 16 2021
%t z = 50; f[x_] := f[x] = Sum[Floor[(3/2)*(k + 1)] x^k, {k, 0, z}]; f[x]
%t CoefficientList[Series[1/f[x], {x, 0, z}], x]
%t LinearRecurrence[{-2},{1,-3,5,-9},40] (* _Harvey P. Dale_, Jul 28 2023 *)
%Y Cf. A005010.
%K sign,easy
%O 0,2
%A _Clark Kimberling_, Dec 18 2016