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

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

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

(If there were a non-trivial integer solution, then u0 �= 0, <strong>and</strong> we obtain a contradiction<br />

to (i).) Thus the parallelotope<br />

�<br />

P = x ∈ E d+1 �<br />

�<br />

: � τ 1 d<br />

α (x1<br />

� �<br />

� �<br />

− ϑ1x0) � ≤ 1,..., � τ 1 d<br />

α (xd<br />

� �<br />

� �<br />

− ϑd x0) � ≤ 1, � 1<br />

in E d+1 contains only the point o of Z d+1 , <strong>and</strong> therefore,<br />

(4) λP = λ1(P, Z d+1 )>1.<br />

Let A be the coefficient matrix of the linear forms which determine P. Then<br />

A −T ⎛<br />

1<br />

0 ... 0<br />

⎜ τ<br />

⎜ −<br />

= ⎜<br />

⎝<br />

τ 1 d ϑ1 τ<br />

α<br />

1 d<br />

... 0<br />

α<br />

...................<br />

− τ 1 d ϑd<br />

0 ...<br />

α<br />

τ 1 ⎞−T<br />

⎛<br />

⎟ τ τϑ1 ... τϑd<br />

⎟ ⎜<br />

⎟ ⎜<br />

⎟ ⎜<br />

⎟ ⎜ 0<br />

⎟ = ⎜<br />

⎟ ⎜<br />

⎟ ⎜<br />

⎟ ⎝<br />

⎠<br />

d<br />

α<br />

α<br />

τ 1 ... 0<br />

d<br />

..............<br />

0 0 ... α<br />

τ 1 ⎞<br />

⎟ .<br />

⎟<br />

⎠<br />

d<br />

τ x0<br />

� �<br />

�<br />

� ≤ 1<br />

Consider the parallelotope<br />

�<br />

Q = x ∈ E d+1 �<br />

�<br />

: � α<br />

τ 1 � �<br />

� �<br />

x1�<br />

≤ 1,..., �<br />

d<br />

α<br />

τ 1 �<br />

�<br />

�<br />

xd�<br />

≤ 1, τ|ϑ1x1 +···+ϑd xd − x0| ≤1<br />

d<br />

�<br />

= x ∈ E d+1 :|x1| ≤ τ 1 d<br />

α ,...,|xd| ≤ τ 1 d<br />

α , |ϑ1x1 +···+ϑd xd − x0| ≤ 1<br />

�<br />

.<br />

τ<br />

Note that A −T is the coefficient matrix of the linear forms which determine Q. Let<br />

λQ = λ1(Q, Z d+1 ). An application of (4) <strong>and</strong> Lemma 23.1 then shows that<br />

1 <br />

(d + 1)| det A| d + 1 .<br />

|x1| ≤ αd−1τ 1 d<br />

d + 1 ,...,|xd| ≤ αd−1τ 1 d<br />

d + 1 , |ϑ1x1<br />

α<br />

+···+ϑd xd − x0| ≤<br />

d<br />

τ(d + 1)<br />

has no non-trivial integer solution <strong>and</strong> thus, a fortiori, no integer solution (u0,...,ud)<br />

where (u1,...,ud) �= o. Put σ = αd−1τ 1 d /(d + 1) <strong>and</strong> β = αd2 /(d + 1) d+1 . Then<br />

the system of inequalities<br />

|x1| ≤σ,...,|xd| ≤σ, |ϑ1x1 +···+ϑd xd − x0| ≤ β<br />

σ d<br />

has no integer solution (u0,...,ud) where (u1,...,ud) �= o. This implies (ii).<br />

(ii)⇒(i) This implication is shown similarly. ⊓⊔

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

Saved successfully!

Ooh no, something went wrong!