%I #56 Nov 30 2022 07:30:02
%S 1,3,1,7,1,3,1,8,1,3,1,7,1,3,1,9,1,3,1,7,1,3,1,8,1,3,1,7,1,3,1,11,1,3,
%T 1,7,1,3,1,8,1,3,1,7,1,3,1,9,1,3,1,7,1,3,1,8,1,3,1,7,1,3,1,15,1,3,1,7,
%U 1,3,1,8,1,3,1,7,1,3,1,9,1,3,1,7,1,3,1,8,1,3,1,7,1,3,1,11,1,3,1,7,1,3
%N Maximal number of linearly independent smooth nowhere-zero vector fields on a (2n+1)-sphere.
%C The corresponding terms for a 2n-sphere are all 0 ("you can't comb the hair on a billiard ball"). The "3" and "7" come from the quaternions and octonions.
%C b(n) = a(n-1): b(2^e) = ((e+1) idiv 4) + 2^((e+1) mod 4) - 1, b(p^e) = 1, p>2. - _Christian G. Bower_, May 18 2005
%C a(n-1) is multiplicative. - _Christian G. Bower_, Jun 03 2005
%H T. D. Noe, <a href="/A053381/b053381.txt">Table of n, a(n) for n = 0..10000</a>
%H J. Frank Adams, <a href="http://dx.doi.org/10.1016/0040-9383(62)90096-4">Vector fields on spheres</a>, Topology, 1 (1962), 63-65.
%H J. Frank Adams, <a href="https://doi.org/10.1090/S0002-9904-1962-10693-4">Vector fields on spheres</a>, Bull. Amer. Math. Soc. 68 (1962) 39-41.
%H J. Frank Adams, <a href="http://www.jstor.org/stable/1970213">Vector fields on spheres</a>, Annals of Math. 75 (1962) 603-632.
%H A. Hurwitz, <a href="http://gdz.sub.uni-goettingen.de/dms/resolveppn/?PPN=GDZPPN002498200">Uber die Komposition der quadratischen formen</a>, Math. Annalen 88 (1923) 1-25.
%H M. Kervaire, <a href="https://dx.doi.org/10.1073/pnas.44.3.280">Non-parallelizability of the sphere for n > 7</a>, Proc. Nat. Acad. Sci. USA 44 (1958) 280-283.
%H J. Milnor, <a href="http://www.jstor.org/stable/1970255">Some consequences of a theorem of Bott</a>, Annals Math. 68 (1958) 444-449.
%H J. Radon, <a href="https://www.semanticscholar.org/paper/Lineare-scharen-orthogonaler-matrizen-Radon/3b8243a9ccb99cd34e3b3764792b8741b45889db">Lineare Scharen Orthogonaler Matrizen</a>, Abh. Math. Sem. Univ. Hamburg 1 (1922) 1-14.
%F Let f(n) be the number of linearly independent smooth nowhere-zero vector fields on an n-sphere. Then f(n) = 2^c + 8d - 1 where n+1 = (2a+1) 2^b and b = c+4d and 0 <= c <= 3. f(n) = 0 if n is even.
%F a((2*n+1)*2^p-1) = A047530(p+1), p >= 0 and n >= 0. a(2*n) = 1, n >= 0, and a(2^p-1) = A047530(p+1), p >= 0. - _Johannes W. Meijer_, Jun 07 2011
%F a(n) = A209675(n+1) - 1. - _Reinhard Zumkeller_, Mar 11 2012
%F Asymptotic mean: Limit_{m->oo} (1/m) * Sum_{k=1..m} a(k) = 10/3. - _Amiram Eldar_, Nov 29 2022
%p with(numtheory): for n from 1 to 601 by 2 do c := irem(ifactors(n+1)[2,1,2],4): d := iquo(ifactors(n+1)[2,1,2],4): printf(`%d,`, 2^c+8*d-1) od:
%p nmax:=101: A047530 := proc(n): ceil(n/4) + 2*ceil((n-1)/4) + 4*ceil((n-2)/4) + ceil((n-3)/4) end: for p from 0 to ceil(simplify(log[2](nmax))) do for n from 0 to ceil(nmax/(p+2))+1 do A053381((2*n+1)*2^p-1) := A047530(p+1): od: od: seq(A053381(n), n=0..nmax); # _Johannes W. Meijer_, Jun 07 2011, revised Jan 29 2013
%t a[n_] := Module[{b, c, d, rho, n0}, n0 = 2*n; b = 0; While[BitAnd[n0, 1] == 0, n0 /= 2; b++]; c = BitAnd[b, 3]; d = (b - c)/4; rho = 2^c + 8*d; Return[rho - 1]]; Table[a[n], {n, 1, 102}] (* _Jean-François Alcover_, May 16 2013, translated from C *)
%o (C) int MaxLinInd(int n){ /* Returns max # linearly indep smooth nowhere zero * vector fields on S^{n-1}, n=1,2,... */ int b,c,d,rho; b = 0; while((n & 1)==0){ n /= 2; b++; } c = b & 3; d = (b - c)/4; rho = (1 << c) + 8*d; return( rho - 1); }
%o (C, gcc/clang) int MaxLinInd(int n) { int b = __builtin_ctz(n); return (1<<b%4) + b/4*8 - 1; } /* _Jeremy Tan_, Apr 09 2021 */
%Y For another version see A003484. Cf. A189995, A001676.
%Y Cf. A047530, A220466.
%K nonn,nice,easy,mult
%O 0,2
%A _Warren D. Smith_, Jan 06 2000
%E More terms from _James A. Sellers_, Jun 01 2000