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”).

Construct sequences P,Q,R by the rules: Q = first differences of P, R = second differences of P, P starts with 1,5,11, Q starts with 4,6, R starts with 2; at each stage the smallest number not yet present in P,Q,R is appended to R; every number appears exactly once in the union of P,Q,R. Sequence gives P.
8

%I #34 Mar 05 2023 03:08:25

%S 1,5,11,20,36,60,94,140,199,272,360,465,588,730,893,1078,1286,1519,

%T 1778,2064,2378,2721,3094,3498,3934,4403,4907,5448,6027,6645,7303,

%U 8002,8743,9527,10355,11228

%N Construct sequences P,Q,R by the rules: Q = first differences of P, R = second differences of P, P starts with 1,5,11, Q starts with 4,6, R starts with 2; at each stage the smallest number not yet present in P,Q,R is appended to R; every number appears exactly once in the union of P,Q,R. Sequence gives P.

%C P can be extended for 10^6 terms, but it is not known if P,Q,R can be extended to infinity.

%C A probabilistic argument suggests that P, Q, R are infinite. - _N. J. A. Sloane_, May 19 2013

%C Martin Gardner (see reference) states that no such triple P,Q,R of sequences exists if it is required that P(1)<Q(1)<R(1).

%D M. Gardner, Weird Numbers from Titan, Isaac Asimov's Science Fiction Magazine, Vol. 4, No. 5, May 1980, pp. 42ff.

%H Christopher Carl Heckman, <a href="/A225376/b225376.txt">Table of n, a(n) for n = 1..10002</a>

%e The initial terms of P, Q, R are:

%e 1 5 11 20 36 60 94 140 199 272 360

%e 4 6 9 16 24 34 46 59 73 88

%e 2 3 7 8 10 12 13 14 15

%p Hofstadter2 := proc (N) local h, dh, ddh, S, lbmex, i:

%p h := 1, 5, 11: dh := 4, 6: ddh := 2:

%p lbmex := 3: S := {h,dh,ddh}:

%p for i from 4 to N do:

%p while lbmex in S do: S := S minus {lbmex}: lbmex := lbmex + 1: od:

%p ddh := ddh, lbmex:

%p dh := dh, dh[-1] + lbmex:

%p h := h, h[-1] + dh[-1]:

%p S := S union {h[-1], dh[-1], ddh[-1]}:

%p lbmex := lbmex + 1:

%p od:

%p if {h} intersect {dh} <> {} then: return NULL:

%p elif {h} intersect {ddh} <> {} then: return NULL:

%p elif {ddh} intersect {dh} <> {} then: return NULL:

%p else: return [h]: fi:

%p end proc: # _Christopher Carl Heckman_, May 12 2013

%t Hofstadter2[N_] := Module[{P, Q, R, S, k, i}, P = {1, 5, 11}; Q = {4, 6}; R = {2}; k = 3; S = Join[P, Q, R]; For[i = 4, i <= N, i++, While[MemberQ[S, k], S = S~Complement~{k}; k++]; AppendTo[R, k]; AppendTo[Q, Q[[-1]] + k]; AppendTo[P, P[[-1]] + Q[[-1]]]; S = S~Union~{P[[-1]], Q[[-1]], R[[-1]]}; k++]; Which[P~Intersection~Q != {}, Return@Nothing, {P}~Intersection~R != {}, Return@Nothing, R~Intersection~Q != {}, Return@Nothing, True, Return@P]];

%t Hofstadter2[36] (* _Jean-François Alcover_, Mar 05 2023, after _Christopher Carl Heckman_'s Maple code *)

%Y Cf. A225377, A225378, A005228, A030124, A037257.

%K nonn

%O 1,2

%A _N. J. A. Sloane_, May 12 2013, based on email from _Christopher Carl Heckman_, May 06 2013

%E Corrected and edited by _Christopher Carl Heckman_, May 12 2013