A generalization of Löwner-John's ellipsoid theorem

preprint2014arXivOpen access

Abstract

We address the following generalization PP of the Lowner-John ellipsoid problem. Given a (non necessarily convex) compact set K⊂RnK\subset R^n and an even integer dd, find an homogeneous polynomial gg of degree dd such that K⊂G:={x:g(x)≤1}K\subset G:=\{x:g(x)\leq1\} and GG has minimum volume among all such sets. We show that PP is a convex optimization problem even if neither KK nor GG are convex! We next show that PP has a unique optimal solution and a characterization with at most (n+d−1d){n+d-1\choose d} contacts points in K∩GK\cap G is also provided. This is the analogue for $d\textgreater{}2$ of the Lowner-John's theorem in the quadratic case d=2d=2, but importantly, we neither require the set KK nor the sublevel set GG to be convex. More generally, there is also an homogeneous polynomial gg of even degree dd and a point a∈Rna\in R^n such that K⊂G_a:={x:g(x−a)≤1}K\subset G\_a:=\{x:g(x-a)\leq1\} and G_aG\_a has minimum volume among all such sets (but uniqueness is not guaranteed). Finally, we also outline a numerical scheme to approximate as closely as desired the optimal value and an optimal solution. It consists of solving a hierarchy of convex optimization problems with strictly convex objective function and Linear Matrix Inequality (LMI) constraints.

Explore this paper’s authors, topics and related work.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.

Reviews 0

Write a reviewWrite

No reviews yet.

Discussion 0

Add a commentWrite

No comments yet.