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.

6 Mixed Volumes <strong>and</strong> Quermassintegrals 103<br />

Hadwiger will be treated, in which the quermassintegrals are characterized as special<br />

valuations.<br />

For more information we refer to Leichtweiss [640], Schneider [905, 908],<br />

Sangwine-Yager [878] <strong>and</strong> McMullen <strong>and</strong> Schneider [716].<br />

Quermassintegrals <strong>and</strong> Intrinsic Volumes<br />

Recall Steiner’s theorem on parallel bodies, Theorem 6.6: Given a convex body C ∈<br />

C,wehave,<br />

V (C + λB d � �<br />

� �<br />

d<br />

d<br />

) = W0(C) + W1(C)λ +···+ Wd(C)λ<br />

1<br />

d<br />

d for λ ≥ 0,<br />

where the coefficients<br />

Wi(C) = V (C,...,C,<br />

B<br />

� �� �<br />

d−i<br />

d ,...,B d<br />

), i = 0,...,d,<br />

� �� �<br />

i<br />

are the quermassintegrals of C.IfC is contained in a subspace of E d , one can define<br />

the quermassintegrals of C both in E d <strong>and</strong> in this subspace. Unfortunately the result<br />

is not the same as the following proposition shows, where v(·) is the volume <strong>and</strong><br />

wi(·), i = 0,...,d − 1, are the quermassintegrals in d − 1 dimensions, κ0 = 1, <strong>and</strong><br />

κi, i = 1,...,d, is the i-dimensional volume of B i .<br />

Proposition 6.7. Let C ∈ C(E d−1 ) <strong>and</strong> embed E d−1 into E d as usual (first d − 1<br />

coordinates). Then<br />

Proof. Let u = (0,...,0, 1). Then<br />

d�<br />

i=0<br />

� d<br />

i<br />

=<br />

i κi<br />

Wi(C) = wi−1(C) for i = 1,...,d.<br />

d κi−1<br />

�<br />

Wi(C)λ i = V (C + λB d ) =<br />

�λ<br />

−λ<br />

i=0<br />

�λ<br />

−λ<br />

v � C + (λ 2 − t 2 ) 1 2 B d−1� dt =<br />

−λ<br />

v � (C + λB d ) ∩ (E d−1 + tu) � dt<br />

�λ<br />

−λ<br />

�<br />

� �<br />

d − 1<br />

wi(C)(λ<br />

i<br />

2 − t 2 ) i �<br />

2 dt<br />

� d−1<br />

�d−1<br />

� � �λ<br />

d − 1<br />

=<br />

wi(C) (λ<br />

i<br />

2 − t 2 ) i �d−1<br />

� �<br />

d − 1<br />

2 dt =<br />

wi(C)<br />

i<br />

κi+1<br />

λ<br />

κi<br />

i+1<br />

for λ ≥ 0<br />

by Steiner’s formula in E d , Steiner’s formula in E d−1 <strong>and</strong> Fubini’s theorem which<br />

is used to calculate the (i + 1)-dimensional volume of the ball λB i+1 . Now equate<br />

coefficients. ⊓⊔<br />

i=0<br />

i=0

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

Saved successfully!

Ooh no, something went wrong!