We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:


Current browse context:


Change to browse by:

References & Citations


(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Mathematics > Metric Geometry

Title: Norms of weighted sums of log-concave random vectors

Abstract: Let $C$ and $K$ be centrally symmetric convex bodies of volume $1$ in ${\mathbb R}^n$. We provide upper bounds for the multi-integral expression \begin{equation*}\|{\bf t}\|_{C^s,K}=\int_{C}\cdots\int_{C}\Big\|\sum_{j=1}^st_jx_j\Big\|_K\,dx_1\cdots dx_s\end{equation*} in the case where $C$ is isotropic. Our approach provides an alternative proof of the sharp lower bound, due to Gluskin and V. Milman, for this quantity. We also present some applications to "randomized" vector balancing problems.
Comments: accepted for publication in Communications in Contemporary Mathematics
Subjects: Metric Geometry (math.MG); Functional Analysis (math.FA); Probability (math.PR)
Cite as: arXiv:1906.03719 [math.MG]
  (or arXiv:1906.03719v1 [math.MG] for this version)

Submission history

From: Giorgos Chasapis [view email]
[v1] Sun, 9 Jun 2019 21:44:21 GMT (21kb)

Link back to: arXiv, form interface, contact.