6
$\begingroup$

I have two questions about the Löwner-John ellipsoid, one just terminology, the other more substantive. Let $K$ be a convex body in $\mathbb{R}^d$.

Q1. Is "the Löwner-John ellipsoid" the unique ellipsoid of maximal volume contained in $K$, or the unique ellipsoid of minimal volume containing $K$? I have seen it used in both senses.

Q2. Let $E^+$ be the containing/circumscribing ellipsoid and $E^-$ the contained/inscribed ellipsoid of min and max volume respectively, for the same $K$. (a) Are there bounds known on $\mathrm{vol}(E^+)/\mathrm{vol}(E^-)$? (b) Any other interesting relationships known between $E^+$ and $E^-$, e.g., alignment of axes?

Thanks for pointers!

$\endgroup$
4
  • 2
    $\begingroup$ All the positive results on Q.2.b that I know are derived from the fact that the ellipsoids are invariant under symmetries of $E$. $\endgroup$ Jan 21, 2012 at 23:24
  • $\begingroup$ @Bill: Thanks, an insightful remark! $\endgroup$ Jan 22, 2012 at 1:30
  • 1
    $\begingroup$ Joseph: You didn't indicate whether you are also interested in sharper results which hold under stronger hypotheses on $K$. If so, there are relevant results in the literature on Banach-Mazur distance. $\endgroup$ Jan 22, 2012 at 14:26
  • 1
    $\begingroup$ I know this is a very old question, but let me just say that $E^+$ is the polar body of $E^-$ of $K^\circ$ (the polar body of $K$). $\endgroup$ Nov 24, 2013 at 17:23

1 Answer 1

10
$\begingroup$

Q1: Most often it is the maximal volume ellipsoid contained in $K$.

Q2: (a) John's theorem implies that $E^+ \subseteq d E^-$ in general, and $E^+ \subseteq \sqrt{d} E^-$ if $K$ is centrally symmetric, and both these inclusions are sharp (consider a simplex or a cube, respectively), giving of course $d^d$ and $d^{d/2}$ for the best possible upper bounds on the ratios of volumes in the nonsymmetric and symmetric cases respectively.

(b) Not that I'm aware of, but there certainly may be some.

The friendliest reference I can think of is K. Ball, "An elementary introduction to modern convex geometry".

$\endgroup$
1
  • $\begingroup$ @Mark: Thanks so much! I should have known of Ball's intro. Downloaded now. :-) $\endgroup$ Jan 21, 2012 at 22:14

Your Answer

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

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