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.

424 <strong>Geometry</strong> of Numbers<br />

lattice packing. Together with the original lattice packing of B d , it forms a packing<br />

of translates of B d with density 2δL(B d ). In this dimension we then have,<br />

δL(B d ) ≤ 1<br />

2 δT (B d ).<br />

29.2 Density Bounds of Blichfeldt <strong>and</strong> Minkowski–Hlawka<br />

While the problem of finding tight lower <strong>and</strong> upper bounds or asymptotic formulae<br />

as d → ∞ for the maximum density of lattice <strong>and</strong> non-lattice packings of balls<br />

remains unsolved, there are substantial pertinent results.<br />

In this section we give Blichfeldt’s upper estimate for δT (B d ) <strong>and</strong> the lower<br />

estimate for δL(B d ) which follows from the Minkowski–Hlawka theorem. Then<br />

more precise upper estimates for δT (B d ) due to Sidel’nikov, Kabat’janski <strong>and</strong><br />

Levenstein <strong>and</strong> lower estimates for δL(B d ) of Schmidt <strong>and</strong> Ball are described.<br />

Upper Estimate for δT (B d ); Blichfeldt’s Enlargement Method<br />

The following result of Blichfeldt [131] was the first substantial improvement of the<br />

trivial estimate δT (B d ) ≤ 1.<br />

Theorem 29.2. δT (B d ) ≤<br />

d + 2<br />

2<br />

2 − d 2 = 2 − d 2 +o(d) as d →∞.<br />

Proof. Thefirststepistoshowtheinequality of Blichfeldt:<br />

(1)<br />

n�<br />

�t j − tk�2 ≤ 2n n�<br />

�s − t j�2 for all t1,...,tn, s ∈ Ed .<br />

Clearly,<br />

j,k=1<br />

j=1<br />

n�<br />

(a j − ak) 2 = �<br />

(a 2 j + a2 k − 2a jak) = �<br />

(a 2 j + a2 � �<br />

k ) − 2<br />

j,k=1<br />

j,k<br />

≤ �<br />

(a 2 j + a2 �<br />

k ) = 2n a 2 j for all a1,...,an ∈ R.<br />

j,k<br />

j<br />

j,k<br />

j<br />

�2 a j<br />

Then,<br />

�<br />

�t j − tk� 2 = �<br />

�(s − t j) − (s − tk)� 2 = � � �<br />

(si − t ji) − (si − tki) �2 j,k<br />

j,k<br />

≤ � � �<br />

(si − t ji) − (si − tki) � � �<br />

2<br />

≤ 2n (si − t ji) 2<br />

i<br />

j,k<br />

= 2n � �<br />

(si − t ji) 2 = 2n �<br />

�s − t j� 2 ,<br />

j<br />

i<br />

j<br />

concluding the proof of Blichfeldt’s inequality (1).<br />

In the second step, we prove the following estimate, where the function f is<br />

defined by f (r) = max � 0, 1 − 1 2 r 2� for r ≥ 0.<br />

j,k<br />

i<br />

i<br />

j

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

Saved successfully!

Ooh no, something went wrong!