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 #17 Mar 11 2023 08:42:21
%S 129,12759,5134241,67898772,11146309948
%N Anti-Waring numbers: least number k such that k and all larger numbers can be expressed as the sum of one or more distinct n-th powers.
%C Sprague finds a(2) in 1948 and proves that a(n) exists for all n >= 2 in the same year. Graham finds a(3) in 1964 with a paper "to appear" with details; Dressler & Parker give an independent proof in 1974. Lin finds a(4) in 1970. Patterson finds a(5) in 1992. Fuller & Nichols, Jr. find a(6) in 2020.
%D S. Lin, Computer experiments on sequences which form integral bases, in J. Leech, ed., Computational Problems in Abstract Algebra, Pergamon Press, 1970, pp. 365-370.
%H R. Dressler and T. Parker, <a href="https://www.ams.org/journals/mcom/1974-28-125/S0025-5718-1974-0327652-1/">12,758</a>, Mathematics of Computation 28:125 (1974), pp. 313-314.
%H Chris Fuller and Robert H. Nichols, Jr., <a href="https://cs.uwaterloo.ca/journals/JIS/VOL18/Fuller/fuller2.html">Generalized anti-Waring numbers</a>, Journal of Integer Sequences 18 (2015), Article 15.10.5.
%H R. L. Graham, <a href="http://www.math.ucsd.edu/~ronspubs/64_04_polynomial.pdf">Complete sequences of polynomial values</a>, Duke Math. J. 31 (1964), pp. 275-285.
%H C. Patterson, <a href="http://hdl.handle.net/1880/46539">The Derivation of a High Speed Sieve Device</a>, Ph.D. thesis, University of Calgary, 1992. [See 2.2.3.2, Complete Sequences, pp. 18-23.]
%H R. Sprague, <a href="https://eudml.org/doc/169077">Über Zerlegung in ungleiche Quadratzahlen</a>, Mathematische Zeitschrift 51 (1948), pp. 289-290.
%H R. Sprague, <a href="https://eudml.org/doc/169088">Über Zerlegungen in n-te Potenzen mit lauter verschiedenen Grundzahlen</a>, Mathematische Zeitschrift, 51 (1948), pp. 466-468.
%F a(n) = A001661(n) + 1. - _Ilya Gutkovskiy_, Mar 24 2022
%e 129 = 2^2 + 5^2 + 10^2, but no subset of {1^2, 2^2, ..., 11^2} sums to 128, so a(2) >= 129.
%e a(3) = 5^3 + 6^3 + 7^3 + 11^3 + 14^3 + 20^3, but a(3) - 1 = 12758 cannot be so represented.
%e a(4) = 2^4 + 6^4 + 7^4 + 14^4 + 28^4 + 46^4
%e a(5) = 2^5 + 3^5 + 6^5 + 8^5 + 9^5 + 10^5 + 13^5 + 14^5 + 19^5 + 22^5 + 27^5 + 29^5 + 30^5
%o (PARI) sumOf(n,k,e,xmax=n)=my(t); if(k==1, my(t); if(ispower(n,e,&t) && t<=xmax, return([t]), return(0))); xmax=min(sqrtnint(n,e),xmax); forstep(x=xmax,k,-1, t=sumOf(n-x^e,k-1,e,x-1); if(t, return(concat(t,x)))); 0
%o bestPowerRep(n,e)=my(k,t); while((t=sumOf(n,k++,e))==0,); t \\ Finds a representation for n as a sum of distinct e-th powers; _Charles R Greathouse IV_, May 04 2020
%Y Cf. A001661.
%K nonn,hard,more
%O 2,1
%A _Charles R Greathouse IV_, Apr 24 2020