9
$\begingroup$

Waring's problem: for fixed $k, s$ write natural numbers $n$ in the form $n=x_1^k+\dots+x_s^k$. $$ G(k):= \min\{s:\text{ all sufficiently large $n$ can be written as above}\} $$ A quick search / Wikipedia / Vaughan-Wooley's survey paper all suggest that the best upper bound for $G(k)$ is (Wooley) $$ G(k)\leq k\log k+k\log\log k+O(k). $$

Recent work of Bourgain, Demeter, Guth (https://arxiv.org/abs/1512.01565) and Wooley (https://arxiv.org/abs/1708.01220) on Vinogradov's theorem seems to improve several quantities related to Waring's problem and similar questions (eg. $g(k)$, which is as $G(k)$ but for all $n$). But I can't find any evidence of improvements in $G(k)$ following from this result. Does anyone know if such improvements are possible, and if so where to find them?

$\endgroup$
  • 5
    $\begingroup$ The breakthroughs on Vinogradov's mean value theorem don't seem to give improvements on $G(k)$. The best results here remain Wooley's work from the 90's giving results of the form $G(k) \le (1+o(1)) k\log k$ (and refinements on making the $o(1)$ here explicit). $\endgroup$ – Lucia Nov 2 '19 at 16:24
  • $\begingroup$ @Lucia Thank you! $\endgroup$ – DaveT Nov 2 '19 at 17:50
4
$\begingroup$

While not exactly the classical quantity $G(k)$, there are partial improvements on asymptotic behavior for the Waring-Goldbach problem, namely for $H(k)$, defined to be the least integer $s$ such that every sufficiently large positive integer congruent to $s$ modulo $K(k)$ may be written as $$p_1^k + p^k_2 + ··· + p^k _s = n,$$ where $p_1,\dots,p_s$ are prime numbers and where $K(k)=\Pi_{(p-1)|k} p^{\gamma}$, and $\gamma$ is a parameter defined below. Shortly after Bourgain, Demeter, and Guth's resolving of the Vinogradov's conjecture Kumcev and Wooley used their results to improve Hua's estimate

$$H(k)\leq k(4 \log k + 2 \log \log k + O(1)) \text{ as }k\rightarrow\infty, \text{ to }$$

$$H(k)\leq (4k − 2) \log k + k − 4 \text{ for } k\geq 3.$$

This was improved again later by Kumcev and Wooley to

$$H(k) ≤ (4k − 2)\log k − (2 \log 2 − 1)k − 3$$

holding for large $k$. To compute the value $K(k)$, let natural $k$ and prime number $p$ be given, and define $\theta = \theta(k, p)$ to be the integer with $p^\theta|k$ but $p^{\theta+1}\nmid k$, and $\gamma=\gamma(k, p)$ by $$ γ(k, p) =\begin{cases} \theta + 2,\ \text{ when } p = 2 \text{ and }\theta > 0,\\ \theta + 1, \text{ otherwise.} \end{cases}$$

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.