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 #21 Feb 07 2023 12:43:46
%S 1,3,2,5,4,7,6,11,8,9,10,13,12,17,14,15,16,19,18,23,20,21,22,25,24,29,
%T 26,27,28,31,30,37,32,33,34,35,36,41,38,39,40,43,42,47,44,45,46,49,48,
%U 53,50,51,52,55,54,59,56,57,58,61,60,67,62,63,64,65,66,71,68
%N Smallest number which is coprime to n and to a(n-1), and is not yet in the sequence; a(1)=1.
%C Permutation of natural numbers with inverse A085230: a(A085230(n))=A085230(a(n))=n.
%C From _Michael De Vlieger_, Apr 13 2022: (Start)
%C Theorem 1: 2 | a(2k+1) for k > 0, consequence of the lexically earliest and coprimality axioms. Even numbers appear in order as a consequence of the latter axiom and since numbers are either even or odd.
%C Corollary: the only fixed point is a(1) = 1.
%C Theorem 2: Generally, if prime p | a(n) then p is coprime to a(n +/- 1). For p = 2, 2 | a(2k+1) for k > 0 since 2 is the smallest prime. For odd p it is not necessarily true that given p | a(n) -> p | a(n+2) or p | a(n-2), since there may be a smaller m such that (a(n-1), m) = 1, q | m for prime q < p, and is not in a(1..n-1).
%C For these reasons, if we also set a(2) = 3, then we need not also check (n, a(n)) = 1, since it isn't possible. If we do not check (n, a(n)) = 1 and set a(2) = 3, 2 would follow 1 since 1 is coprime to all numbers.
%C Theorem 3: 3 | a(3k+1) for k > 1. Proof: For even k, 6 | a(3k+1), i.e., 6 | a(n) : n mod 6 = 1, and it is easy to see that since even numbers appear in order in the sequence, these even multiples of 3 are also in order. Because 3 | a(n) : n mod 6 = 1, we cannot have 3 | a(n) for n congruent to 0 or 2 (mod 6). Furthermore, we know that 2 | a(n) for n congruent to 3 or 5 (mod 6). So 3 | a(n) odd : n mod 6 = 4, that is, 3 | a(3k+1) for k > 1.
%C Theorem 4: Odd primes q set records. Proof: (q, a(n-1)) = 1 as a consequence of lexically earliest axiom that rules out equality, and by the definition of prime. 2 is displaced on account of the axiom that bans equality between n and a(n). Therefore, whereupon q is the smallest unused odd number, it enters the sequence.
%C A consequence of theorems 1 and 3 is that powers of 2 and those of 3 excepting 3 itself do not set records, since their adjacency is governed by a(n-1). The powers of other primes do set records since coprimality does not depend on multiplicity.
%C The smallest composite record is a(24) = 25. Smallest record m with omega(m) > 1 is a(54) = 55. Powers of 2 and 3 are absent from records for n <= 2^20. (End)
%H Michael De Vlieger, <a href="/A085229/b085229.txt">Table of n, a(n) for n = 1..10000</a>
%H Michael De Vlieger, <a href="/A085229/a085229.png">Annotated log-log scatterplot of a(n)</a> for n = 1..256, showing records in red, local minima in blue, and highlighting fixed points in amber and primes in green.
%H Amit Kumar Basistha and Eugen J. Ionascu, <a href="https://arxiv.org/abs/2302.02838">A special sequence and primorial numbers</a>, arXiv:2302.02838 [math.NT], 2023.
%H <a href="/index/Per#IntegerPermutation">Index entries for sequences that are permutations of the natural numbers</a>
%t nn = 10000; c[_] = 0; a[1] = c[1] = 1; a[2] = 3; c[3] = u = 2; Do[k = u; While[Nand[c[k] == 0, CoprimeQ[a[i-1], k]], k++]; Set[{a[i], c[k]}, {k, i}]; If[a[i] == u, While[c[u] > 0, u++]], {i, 3, nn}]; Array[a, nn] (* _Michael De Vlieger_, Apr 13 2022 *)
%o (Python)
%o from math import gcd
%o from itertools import count, islice
%o def agen(): # generator of terms
%o an, aset = 1, {1}
%o for n in count(2):
%o yield an
%o k = 1
%o while k in aset or any(gcd(t, k) != 1 for t in [n, an]): k+= 1
%o aset.add(k)
%o an = k
%o print(list(islice(agen(), 69))) # _Michael S. Branicky_, Apr 13 2022
%Y Cf. A079578, A084937, A093714.
%K nonn
%O 1,2
%A _Reinhard Zumkeller_, Jun 22 2003