14.02.2013 Views

Gruber P. Convex and Discrete Geometry

Gruber P. Convex and Discrete Geometry

Gruber P. Convex and Discrete Geometry

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

8 The Brunn–Minkowski Inequality 165<br />

Using the definition of Minkowski surface area, this inequality readily yields the<br />

isoperimetric inequality. Expressed otherwise, (1) says that the volume of the εneighbourhood<br />

Cε = C + εB d of the compact set C is at least as large as the<br />

volume of the ε-neighbourhood of a (solid Euclidean) ball of the same volume as C.<br />

This result may be considered as a sort of isoperimetric inequality, where the notion<br />

of surface area is not needed. This is the starting point for general isoperimetric<br />

inequalities in metric probability spaces.<br />

The Problem on Metric Probability Spaces<br />

Let 〈M,δ,µ〉 be a space M withametricδ <strong>and</strong> thus a topology <strong>and</strong> a Borel probability<br />

measure µ.Bytheε-neighbourhood Cε of a compact set C in M we mean the<br />

set<br />

Cε ={x ∈ M : δC(x) ≤ ε}, where δC(x) = dist(x, C) = min{δ(x, y) : y ∈ C}.<br />

Then the following general isoperimetric or Brunn–Minkowski problem arises:<br />

Problem 8.2. Given α, ε > 0, for what compact sets C in M with µ(C) = α has the<br />

ε-neighbourhood Cε of C minimum measure?<br />

The Spherical Case<br />

Let S d−1 be endowed with its chordal metric, i.e. the metric inherited from E d , <strong>and</strong><br />

let S be the normalized area measure on S d−1 .Anisoperimetric inequality for S d−1<br />

of Lévy [653], p.269, <strong>and</strong> Schmidt [891] is as follows: Let α > 0. Then, for all<br />

compact sets C ⊆ S d−1 with S(C) = α, wehave,<br />

S(Cε) ≥ S(Kε) for ε ≥ 0,<br />

where K is a spherical cap with S(K ) = α. The equality case was studied by Dinghas<br />

[269].<br />

This result has the following surprising consequence: Let α = 1 2 , so that C has<br />

the measure of a hemisphere H. Then, for each ε ≥ 0, we have that S(Cε) ≥ S(Hε).<br />

An easy proof shows that the complement of Hε has area measure about e− 1 2 dε2 .<br />

Hence S(Cε) is about 1 − e− 1 2 dε2 . Thus, for any given ε>0 <strong>and</strong> large d almost all<br />

of Sd−1 lies within distance ε of any given compact set C in Sd−1 of measure 1 2 .In<br />

other words, for large d the area of Sd−1 is concentrated close to any compact set of<br />

measure 1 2 . As a consequence, for large d the area of Sd−1 is concentrated near any<br />

great circle.<br />

The situation just described becomes even more striking if it is interpreted in<br />

terms of a Lipschitz function f : Sd−1 → R with Lipschitz constant 1. There is<br />

a number m, themedian of f , such that there are compact sets C, D ⊆ Sd−1 with<br />

S(C ∩ D) = 0, S(C) = S(D) = 1 2 <strong>and</strong> C ∪ D = Sd−1 , where f (u) ≤ m for u ∈ C<br />

<strong>and</strong> f (v) ≥ m for v ∈ D. Then S(Cε) <strong>and</strong> S(Dε) both have area about 1 − e− 1 2 dε2 .

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!