A046185 -id:A046185 - cp's OEIS frontend

A005282 Mian-Chowla sequence (a B_2 sequence): a(1) = 1; for n>1, a(n) = smallest number > a(n-1) such that the pairwise sums of elements are all distinct.

1, 2, 4, 8, 13, 21, 31, 45, 66, 81, 97, 123, 148, 182, 204, 252, 290, 361, 401, 475, 565, 593, 662, 775, 822, 916, 970, 1016, 1159, 1312, 1395, 1523, 1572, 1821, 1896, 2029, 2254, 2379, 2510, 2780, 2925, 3155, 3354, 3591, 3797, 3998, 4297, 4433, 4779, 4851

Offset: 1

N. J. A. Sloane and Simon Plouffe

An alternative definition is to start with 1 and then continue with the least number such that all pairwise differences of distinct elements are all distinct. - Jens Voß, Feb 04 2003. [However, compare A003022 and A227590. - N. J. A. Sloane, Apr 08 2016]
Rachel Lewis points out [see link] that S, the sum of the reciprocals of this sequence, satisfies 2.158435 <= S <= 2.158677. Similarly, the sum of the squares of reciprocals of this sequence converges to approximately 1.33853369 and the sum of the cube of reciprocals of this sequence converges to approximately 1.14319352. - Jonathan Vos Post, Nov 21 2004; comment changed by N. J. A. Sloane, Jan 02 2020
Let S denote the reciprocal sum of a(n). Then 2.158452685 <= S <= 2.158532684. - Raffaele Salvia, Jul 19 2014
From Thomas Ordowski, Sep 19 2014: (Start)
Known estimate: n^2/2 + O(n) < a(n) < n^3/6 + O(n^2).
Conjecture: a(n) ~ n^3 / log(n)^2. (End)

			The second term is 2 because the 3 pairwise sums 1+1=2, 1+2=3, 2+2=4 are all distinct.
The third term cannot be 3 because 1+3 = 2+2. But it can be 4, since 1+4=5, 2+4=6, 4+4=8 are distinct and distinct from the earlier sums 1+1=2, 1+2=3, 2+2=4.

P. Erdős and R. Graham, Old and new problems and results in combinatorial number theory. Monographies de L'Enseignement Mathématique (1980).
S. R. Finch, Mathematical Constants, Cambridge, 2003, Section 2.20.2.
R. K. Guy, Unsolved Problems in Number Theory, E28.
A. M. Mian and S. D. Chowla, On the B_2-sequences of Sidon, Proc. Nat. Acad. Sci. India, A14 (1944), 3-4.
N. J. A. Sloane and Simon Plouffe, The Encyclopedia of Integer Sequences, Academic Press, 1995 (includes this sequence).

T. D. Noe, Table of n, a(n) for n=1..5818 (terms less than 2*10^9)
Thomas Bloom, Problem 340, Erdős Problems.
Yin Choi Cheng, Greedy Sidon sets for linear forms, J. Num. Theor. (2024).
Rachel Lewis, Mian-Chowla and B2 sequences, 1999. [Thanks to _Steven Finch_ for providing this document. Included with permission. - _N. J. A. Sloane_, Jan 02 2020]
Kevin O'Bryant, B_h-Sets and Rigidity, arXiv:2312.10910 [math.NT], 2023.
Raffaele Salvia, Table of n, a(n) for n=1...25000
R. Salvia, A New Lower Bound for the Distinct Distance Constant, J. Int. Seq. 18 (2015) # 15.4.8.
N. J. A. Sloane, Handwritten notes on Self-Generating Sequences, 1970 (note that A1148 has now become A005282)
Eric Weisstein's World of Mathematics, B2 Sequence.
Eric Weisstein's World of Mathematics, Chowla Sequence.
Zhang Zhen-Xiang, A B_2-sequence with larger reciprocal sum, Math. Comp. 60 (1993), 835-839.
Index entries for B_2 sequences.

Row 2 of A347570.
Cf. A051788, A080200 (for differences between terms).
Different from A046185. Cf. A011185.
See also A003022, A227590.
A259964 has a greater sum of reciprocals.
Cf. A002858.

import Data.Set (Set, empty, insert, member)
a005282 n = a005282_list !! (n-1)
a005282_list = sMianChowla [] 1 empty where
   sMianChowla :: [Integer] -> Integer -> Set Integer -> [Integer]
   sMianChowla sums z s | s' == empty = sMianChowla sums (z+1) s
                        | otherwise   = z : sMianChowla (z:sums) (z+1) s
      where s' = try (z:sums) s
            try :: [Integer] -> Set Integer -> Set Integer
            try []     s                      = s
            try (x:sums) s | (z+x) `member` s = empty
                           | otherwise        = try sums $ insert (z+x) s
-- Reinhard Zumkeller, Mar 02 2011

a[1]:= 1: P:= {2}: A:= {1}:
for n from 2 to 100 do
  for t from a[n-1]+1 do
    Pt:= map(`+`,A union {t},t);
    if Pt intersect P = {} then break fi
  od:
  a[n]:= t;
  A:= A union {t};
  P:= P union Pt;
od:
seq(a[n],n=1..100); # Robert Israel, Sep 21 2014

t = {1}; sms = {2}; k = 1; Do[k++; While[Intersection[sms, k + t] != {}, k++]; sms = Join[sms, t + k, {2 k}]; AppendTo[t, k], {49}]; t (* T. D. Noe, Mar 02 2011 *)

A005282_vec(N, A=[1], U=[0], D(A, n=#A)=vector(n-1, k, A[n]-A[n-k]))={ while(#A2 && U=U[k-1..-1]);A} \\ M. F. Hasler, Oct 09 2019

aupto(L)= my(S=vector(L), A=[1]); for(i=2, L, for(j=1, #A, if(S[i-A[j]], next(2))); for(j=1, #A, S[i-A[j]]=1); A=concat(A, i)); A \\ Ruud H.G. van Tol, Jun 30 2025

from itertools import count, islice
def A005282_gen(): # generator of terms
    aset1, aset2, alist = set(), set(), []
    for k in count(1):
        bset2 = {k<<1}
        if (k<<1) not in aset2:
            for d in aset1:
                if (m:=d+k) in aset2:
                    break
                bset2.add(m)
            else:
                yield k
                alist.append(k)
                aset1.add(k)
                aset2 |= bset2
A005282_list = list(islice(A005282_gen(),30)) # Chai Wah Wu, Sep 05 2023

a(n) = A025582(n) + 1.

a(n) = (A034757(n)+1)/2.

Examples added by N. J. A. Sloane, Jun 01 2008

A384729 A B_2-sequence with reciprocal sum > 2.1615.

Original entry on oeis.org

1, 2, 4, 8, 13, 21, 31, 45, 66, 81, 97, 123, 148, 182, 204, 252, 291, 324, 352, 415, 486, 540, 651, 706, 792, 838, 928, 1046, 1134, 1228, 1358, 1407, 1512, 1624, 1869, 1938, 2087, 2170, 2367, 2480, 2608, 2765, 3033, 3080, 3232, 3567, 3605, 3797, 3950, 4267, 4505, 4677, 5064, 5290, 5480, 5655, 6059, 6507, 6892, 6967

Offset: 1

Logan J. Kleinwaks, Jun 08 2025

nonn

This is the B_2 sequence with largest reciprocal sum that is known to the author, as of the date of submission. The reciprocal sum of the first 1010 terms, which are given in the attached b-file, is 2.16150003... This already provably exceeds the reciprocal sum of the infinite sequence by R. Lewis (A046185), defined by greedy extension of a sequence of 68 terms (e.g., compute R. Lewis' sequence to 1600 terms, then bound the remainder using the result of B. Lindström on the maximum cardinality of B2-sequences with elements in [1, N]). By extending this sequence greedily after the first 1010 terms, a larger reciprocal sum can obviously be achieved.
This sequence first differs from A046185 at the 25th term.
This sequence was found by the author by using a combination of beam search and batch greedy algorithm, as part of an experiment to evaluate LLM code generation and mathematical reasoning (all code was written by the LLM, though with significant prompting).

Logan J. Kleinwaks, Table of n, a(n) for n = 1..1010
Rachel Lewis, Mian-Chowla and B2 sequences, 1999.
Bernt Lindström, An inequality for B_2-sequences, Journal of Combinatorial Theory, Volume 6, Issue 2 (1969), 211-212.
Eric W. Weisstein, B_2-Sequence.. From MathWorld--A Wolfram Web Resource.

Cf. A046185 (R. Lewis), A005282 (Mian-Chowla).

cp's OEIS Frontend

A005282 Mian-Chowla sequence (a B_2 sequence): a(1) = 1; for n>1, a(n) = smallest number > a(n-1) such that the pairwise sums of elements are all distinct.

Views

Author

Keywords

Comments

Examples

References

Links

Crossrefs

Programs

Haskell

Maple

Mathematica

PARI