The are some linearly upper bounds on Hermite's constant $\gamma_d$, such as $\gamma_d \leq 2d/3$, $\gamma_d \leq d/4+1$. So we can claim that $\gamma_d=O(d)$. There is also a rather tight asymptotical bound for $\gamma_n$: $\frac{d}{2\pi e}+\frac{\log(\pi d)}{2\pi e}+o(1) \leq \gamma_d \leq \frac{1.744d}{2 \pi e}(1+o(1))$ (see page 34 of The LLL algorithm edited by Phong Q. Nguyen et al.). After this tight asymptotical bound, there is a sentence that "Thus, $\gamma_d$ is enssentially linear in $d$" in this book (also in page 34). My questions is that can we express $\gamma_d$ in the asymptotical form of $\Theta(d)$. I have searched some related papers, but I can't find the the corresponding expression.

  • $\begingroup$ What you want seems to follow from the inequality you have linked? The first inequality implies that $\gamma_d = \Omega(d)$. The second inequality implies that $\gamma_d = O(d)$. Combined you have that $\gamma_d =\Theta(d)$. $\endgroup$
    – Mark
    Aug 2, 2023 at 5:39


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.