All integers except possibly those of the form $9k \pm 4$ can be written as the sum of 4 cubes.
Are there any heuristics that suggest this is impossible in general, or is it the case that no one has found any identities?
All integers except possibly those of the form $9k \pm 4$ can be written as the sum of 4 cubes.
Are there any heuristics that suggest this is impossible in general, or is it the case that no one has found any identities?
Copyright © 2021 JogjaFile Inc.
(Updated):
If the number is of the form $9k\pm 4$, then it cannot be written as the sum of three cubes. Simply because any cube number has form $9s$ or $9s\pm 1$.
As for four cubes: maybe it is not proven that all numbers of the form $9k\pm 4$ can be written as the sum of four cubes.
It looks that for many values $k$ it is relatively easy to decompose numbers of the form $9k\pm 4$ to the sum of $4$ cubes.
I checked $k$ in the range $1\le k \le 10^4$, and result is:
if $1\le k \le 10000$, then equation $$9k-4=a^3+b^3+c^3+d^4\tag{1}$$ has integer solution $(k,a,b,c,d)$, where $\max\{|a|,|b|,|c|,|d|\}<300$.
Few relatively "hard" examples from this set:
$9\cdot 3112-4 = 28004 = (-175)^3 + (-214)^3 + (-226)^3 + 299^3$;
$9\cdot 9174-4 = 82562 = 104^3 + (-196)^3 + (-235)^3 + 269^3$;
$9\cdot 1264-4 = 11372 = 59^3 + 59^3 + 257^3 + (-259)^3$;
$9\cdot 1418-4 = 12758 = (-58)^3 + (-169)^3 + (-217)^3 + 248^3$;
$9\cdot 3105-4 = 27941 = (-100)^3 + (-106)^3 + (-232)^3 + 245^3$;
and equation $$9k+4=a^3+b^3+c^3+d^4\tag{2}$$ has integer solution $(k,a,b,c,d)$, where $\max\{|a|,|b|,|c|,|d|\}<520$.
Few relatively "hard" examples from this set:
$9\cdot 2474+4 = 22270 = 325^3 + 349^3 + 391^3 + (-515)^3$;
$9\cdot 6517+4 = 58657 = (-161)^3 + (-221)^3 + (-401)^3 + 430^3$;
$9\cdot 6608+4 = 59476 = (-101)^3 + 247^3 + 295^3 + (-341)^3$;
$9\cdot 7479+4 = 67315 = (-221)^3 + 250^3 + 307^3 + (-323)^3$;
$9\cdot 574+4 = 5170 = 22^3 + 67^3 + 319^3 + (-320)^3$;
Here are few first numbers of the form $9k-4$:
And few first numbers of the form $9k+4$: