cp's OEIS Frontend

This is a front-end for the Online Encyclopedia of Integer Sequences, made by Christian Perfect. The idea is to provide OEIS entries in non-ancient HTML, and then to think about how they're presented visually. The source code is on GitHub.

A014742 Expansion of (1+x^2)/(1 - 2*x - 2*x^2 + x^3).

Original entry on oeis.org

1, 2, 7, 17, 46, 119, 313, 818, 2143, 5609, 14686, 38447, 100657, 263522, 689911, 1806209, 4728718, 12379943, 32411113, 84853394, 222149071, 581593817, 1522632382, 3986303327, 10436277601, 27322529474, 71531310823, 187271402993, 490282898158, 1283577291479
Offset: 0

Views

Author

Keywords

Comments

Let M = a triangle with (1,1,1,3,3,5,5,7,7,...) as the left border and (0,1,2,3,4,5,...) as all other columns. A014742 = lim_{n->infinity} M^n, the left-shifted vector considered as a sequence. - Gary W. Adamson, Jul 26 2010
For n >= 1, a(n) is the ratio of L/h (rounded down), where L = length of short sides of parallelogram appearing in dissection fallacy of square F(n+2) X F(n+2), F(n) is Fibonacci number, and h = perpendicular distance between the long sides LL. The first differences of A069921 give L^2. See illustration. - Kival Ngaokrajang, Jun 27 2014
From Wolfdieter Lang, Jul 15 2014: (Start)
The preceding comment is a conjecture using a(n) = floor(LL(n)*L(n)) with LL(n) = sqrt(F(n+2)^2 + F(n)^2) and L(n) = LL(n-1), n >= 1 (its author agreed with this in an email). See also, e.g., the Koshy reference for the dissection fallacy, sect. 6, 100 - 108.
The proof of the conjecture uses first the identity (LL(n)*LL(n-1))^2 - a(n)^2 = + 1 with a(n) = F(n-1)*F(n) + F(n+1)*F(n+2) (see the formula section for a(n)). This identity is due to the factorization of the left-hand side which is A(n)^2 with A(n) = F(n)*F(n+1) - F(n-1)*F(n+2). But A(n) = (-1)^(n+1) is a special instance of a well known Fibonacci identity (Koshy, p. 88, Nr. 19 for h=-1, k=2, F(-1) = 1). Now one has (LL(n)*LL(n-1))^2 = 1 + a(n)^2, that is LL(n)*LL(n-1) = sqrt(1 + a(n)^2). Because a(n) < sqrt(1 + a(n)^2) < a(n) + 1 (just square both inequalities using a(n) > 0) one has now proved that floor(LL(n)*LL(n-1)) = a(n), n >= 1. (End)
a(n) = numerator(Re(C(n))), with the complex sequence C(n) defined in the name of A069921. - Wolfdieter Lang, Jul 16 2014

Examples

			a(2) = F(1)*F(2) + F(3)*F(4) = 1*1 + 2*3 = 7. - _James R. Buddenhagen_, Jan 06 2009
		

References

  • T. Koshy, Fibonacci and Lucas Numbers with Applications, John Wiley & Sons, 2001.

Crossrefs

Cf. A069921. - Kival Ngaokrajang, Jun 27 2014
Cf. similar sequences of the type k*F(n)*F(n+1) + (-1)^n listed in A264080.

Programs

  • Maple
    seq(combinat[fibonacci](n-1)*combinat[fibonacci](n)+combinat[fibonacci](n+1)*combinat[fibonacci](n+2), n=0..50); # will give first 50 terms - James R. Buddenhagen, Jan 06 2009
  • Mathematica
    CoefficientList[Series[(1 + x^2)/(1 - 2*x - 2*x^2 + x^3), {x, 0, 30}], x] (* Wesley Ivan Hurt, Jun 27 2014 *)
    LinearRecurrence[{2, 2, -1},{1, 2, 7},30] (* Ray Chandler, Sep 23 2015 *)
  • PARI
    Vec((1+x^2)/(1-2*x-2*x^2+x^3)+O(x^99)) \\ Charles R Greathouse IV, Sep 26 2012
    
  • PARI
    a(n) = round((2^(-1-n)*((-1)^n*2^(2+n)-3*(3-sqrt(5))^n*(-1+sqrt(5))+3*(1+sqrt(5))*(3+sqrt(5))^n))/5) \\ Colin Barker, Sep 29 2016

Formula

a(n) = F(n-1)*F(n) + F(n+1)*F(n+2), where F = A000045. - James R. Buddenhagen, Jan 06 2009
From Wolfdieter Lang, Jul 15 2014: (Start)
G.f.: (1+x^2)/(1 - 2*x - 2*x^2 + x^3) = (1+x^2)/((1+x)*(1 - 3*x + x^2)) = (2/(1+x) + 3*(1+x)/(1 - 3*x + x^2))/5 (see the name).
a(n) = (2*(-1)^n + 3*(F(2*n) + F(2*(n+1))))/5,
a(n) = (2*(-1)^n + L(2*n-1) + L(2*n+3))/5 with L(n) = A000032(n) and L(-1) = -1. (End)
a(n) = 3*F(n)*F(n+1) + (-1)^n. - Bruno Berselli, Oct 30 2015
a(n) = (2^(-1-n)*((-1)^n*2^(2+n) - 3*(3-sqrt(5))^n*(-1+sqrt(5)) + 3*(1+sqrt(5))*(3+sqrt(5))^n))/5. - Colin Barker, Sep 29 2016
a(n) = Fibonacci(n + 2)^2 - 2*Fibonacci(n)^2. - Detlef Meya, Jun 29 2024
E.g.f.: exp(-x)*(2 + 3*exp(5*x/2)*(cosh(sqrt(5)*x/2) + sqrt(5)*sinh(sqrt(5)*x/2)))/5. - Stefano Spezia, Jun 30 2024

Extensions

Buddenhagen's Jan 06 2009 entries adjusted for offset 0 by Wolfdieter Lang, Jul 15 2014