Is there a constant $C$ such that for every tetrahedron with edges $a,b,c,d,e,f$ and volume $V$, the following inequality holds: $$abcdef\ge C\cdot V^2$$

The trivial case $a=b=c=d=e=f$ gives $72\ge C$, but this is all I obtained.

This question is a 3D-version of inequality involving the product of lengths of edges of quadrilateral

The method used in RobertZ's solution seems not work here, since both sides would tend to 0.

  • 1
    $\begingroup$ A better analogy in the planar case would be that of a triangle. The corresponding inequality should be $$abc\ge C\cdot A^{3/2}$$ and in that case I think one can prove $C=8/\sqrt[4]{27}$. $\endgroup$ – Aretino Jul 25 '17 at 16:48
  • $\begingroup$ Thanks for your comment. Yes, it's true. This inequality is equivalent to $\frac{1}{2}R\ge\frac{\sqrt{S\sqrt{3}}}{3}$ which is known and can be foung e.g. on imomath.com/index.php?options=604&lmm=0 (I wrote $S$ instead of $A$, because this letter is used on the page I gave link to) $\endgroup$ – tong_nor Jul 25 '17 at 17:38
  • $\begingroup$ By analogy with that case I think $C$ attains its bound when all edges are equal, i.e. $C=72$. $\endgroup$ – Aretino Jul 25 '17 at 17:44

According to a chinese book Guide of Distance Geometric Analysis (距离几何分析导引) on my bookshelf, there is something called Veljan-Korchmáros inequality:

Let $v_1, v_2, \ldots, v_{n+1} \in \mathbb{E}^n$ be the vertices of a $n$-simplex. Let $a_{ij} = |v_i - v_j|$ be the edge lengths and $V$ be the volume of the simplex. We have $$\left(\prod_{1\le i < j \le n+1} a_{ij}\right)^{\frac{2}{n+1}} \ge n!\sqrt{\frac{2^n}{n+1}} V$$ and the equality is true when and only when the simplex is regular.

For $n = 3$, this implies $C = 72$.

I can't find an english reference of this inequality online. Following is the reference quoted in above book.

  • Korchmáros G. Una limitazione per il volume di un simplesso $n$-dimensionale avente spigoli di date lunghezze. Atti Accad. Naz. Lincei Rend. Cl. Sci. Fis. Mat. Natur., 1974, 56(6): 876-879.
  • 2
    $\begingroup$ I corrected the reference, and found here a proof of the theorem. $\endgroup$ – Aretino Jul 25 '17 at 19:11
  • $\begingroup$ @Aretino thanks for the fix, I just copied what is cited on the book which seems to insert spaces at random places. $\endgroup$ – achille hui Jul 25 '17 at 19:47

Let $a=b=c=d=e=1$ and $f=x$.

Hence, $V^2=\frac{x^2(3-x^2)}{144}$.

Thus, $$x\geq\frac{Cx^2(3-x^2)}{144}$$ or $$C\leq\frac{144}{x(3-x^2)}$$ and by AM-GM $$\frac{144}{3x-x^3}=\frac{144}{2-(2-3x+x^3)}\geq\frac{144}{2}=72$$

Thus, $C\leq72$, which is you know alredy.


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.