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.

(5)<br />

�<br />

C<br />

v(x) dx = lim<br />

�<br />

n→∞<br />

C<br />

vn(x) dx = lim<br />

n→∞ αvn(o) = αv(o)<br />

12 Special <strong>Convex</strong> Bodies 225<br />

by Proposition (3). For the notions of stochastic convergence <strong>and</strong> uniform integrability<br />

<strong>and</strong> for the first equality in (5) see, e.g. Bauer [82], Sects. 20, 21.<br />

Finally, let a harmonic function w be defined by<br />

w(x) = 1 +�p�v(x) for x ∈ E 3 \{p}<br />

<strong>and</strong> let B be the ball �p�B 3 . Since o ∈ int C by (2), our choice of p shows that<br />

B ⊆ C. Thus<br />

(6) αw(o) = α � 1 +�p�v(o) � �<br />

= 0 =<br />

�<br />

� �<br />

1 +�p�v(x) dx = w(x) dx<br />

by the definitions of w <strong>and</strong> v, (3) applied to u = 1 <strong>and</strong> Proposition (5). The mean<br />

value property of harmonic functions implies that<br />

�<br />

(7) w(x) dx = γw(o) = 0,<br />

B<br />

where γ > 0 is a suitable multiplicative constant, see Wermer [1018], Appendix.<br />

From (6) <strong>and</strong> (7) we conclude that<br />

�<br />

�<br />

�<br />

�<br />

0 = w(x) dx = w(x) dx + w(x) dx = w(x) dx.<br />

C<br />

B<br />

Since w(x) > 0forx ∈ C\B, it follows that C\B has measure 0. Noting that<br />

C = cl int C by assumption, this means that C\B =∅,orC ⊆ B. The reverse<br />

inclusion being obvious, C = B follows. ⊓⊔<br />

C<br />

C\B<br />

C\B<br />

12.3 Blaschke’s Characterization of Ellipsoids <strong>and</strong> Its Applications<br />

Ellipsoids <strong>and</strong> Euclidean spaces play an important role in many branches of mathematics.<br />

Among these are convex <strong>and</strong> differential geometry, the local theory of<br />

normed spaces, functional analysis, approximation, operator <strong>and</strong> potential theory,<br />

dynamical systems, combinatorial optimization <strong>and</strong> mechanics. The first characterizations<br />

of ellipsoids in convex geometry go back to Brunn <strong>and</strong> Blaschke, <strong>and</strong> at<br />

present there is still interest for these characterizations. One may distinguish between<br />

characterizations based on:<br />

Affine <strong>and</strong> projective transformations<br />

Sections<br />

Projections <strong>and</strong> illuminations<br />

Extremal properties<br />

Other geometric <strong>and</strong> analytic properties<br />

C

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

Saved successfully!

Ooh no, something went wrong!