1,2

a(n)=2a(n-1)-a(n-2)+2a(n-3). Sequence is identical to its half second differences from the second term; a(n)+a(n+2)=2^(n+2). - Paul Curtz (bpcrtz(AT)free.fr), Dec 17 2007

Also describes the location a(n) of the minimal scaling factor when rescaling an FFT of order 2^{n+2} in order to (currently) minimize the arithmetic operation count (Johnson & Frigo, 2007). - Steven G. Johnson (stevenj(AT)math.mit.edu), Dec 27 2006

I. Gessel, Problem 10424, Amer. Math. Monthly, 102 (1995), 70.

M. E. Larsen, Summa Summarum, A. K. Peters, Wellesley, MA, 2007; see p. 38. [From N. J. A. Sloane (njas(AT)research.att.com), Jan 29 2009]

Index entries for sequences related to linear recurrences with constant coefficients

S. G. Johnson and M. Frigo, A modified split-radix FFT with fewer arithmetic operations, IEEE Trans. Signal Processing 55 (2007), 111-119.

a(1) = 1, a(2n+1) = 2*a(2n) and a(2n) = 2*a(2n-1) + (-1)^n.

a(n) = (4*2^n+cos(pi*n/2)+2sin(pi*n/2))/5. - Paul Barry (pbarry(AT)wit.ie), Dec 17 2003

V:=n->(1/5)*(2^(n-1)+2*cos(n*Pi/2)-sin(n*Pi/2)); [seq(V(n), n=0..12)];

Cf. A007909, A007679.

Sequence in context: A018775 A086514 A079662 this_sequence A052702 A058766 A127601

Adjacent sequences: A007907 A007908 A007909 this_sequence A007911 A007912 A007913

nonn,easy

Mogens Esrom Larsen (mel(AT)math.ku.dk)

Entry revised Feb 24 2004 - N. J. A. Sloane (njas(AT)research.att.com).