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.

(t) = adt +···+a1t d + a0t d+1<br />

19 Lattice Polytopes 315<br />

<strong>and</strong> we obtain, from (13)<br />

L o � � � �<br />

� �<br />

d + n − 1 d + n − 2<br />

n − 1<br />

(nS) =<br />

ad +<br />

ad−1 +···+ a0<br />

d<br />

d<br />

d<br />

= 1 � �<br />

(d + n − 1) ···(n + 1)nad +···+(n − 1) ···(n − 1 − d + 1)a0<br />

d !<br />

= (−1)d � �<br />

(−n + d) ···(−n + 1)a0 +···+(−n) ···(−n − d + 1)ad<br />

d !<br />

= (−1) d<br />

�� � � � �<br />

−n + d<br />

−n<br />

a0 +···+ ad = (−1)<br />

d<br />

d<br />

d pS(−n).<br />

This, together with the corresponding results in lower dimensions, yields the following<br />

counterpart of (6) <strong>and</strong> (9).<br />

(14) Let S ∈ P Z d be a simplex with c = dim S ≤ d. Then<br />

L o (nS) = (−1) c pS(−n) for n ∈ N.<br />

In the final step we extend (14) to proper convex lattice polytopes.<br />

(15) Let P ∈ P Z d p . Then L o (nP) = (−1) d pP(−n) for n ∈ N.<br />

Represent P as the disjoint union of the relative interiors of the simplices of a complex<br />

of lattice simplices as in the proof of (12). Then<br />

(16) pP(n) = L(nP) = �<br />

L<br />

S<br />

o (nS) + �<br />

L<br />

R<br />

o (nR) +···<br />

= �<br />

(−1) d pS(−n) + �<br />

(−1) d−1 pR(−n) +··· for n ∈ N,<br />

S<br />

R<br />

where S ranges over all proper simplices of this complex, R over all (d − 1)dimensional<br />

simplices, etc. Since pP, pS,..., all are polynomials <strong>and</strong> since (16)<br />

holds for all n ∈ N, it holds for all t ∈ R in place of n ∈ N. Thus it holds in<br />

particular for all −n where n ∈ N. Hence<br />

(17) (−1) d pP(−n) = �<br />

pS(n) −<br />

S<br />

�<br />

pR(n) +···<br />

R<br />

= �<br />

L(nS) − �<br />

L(nR) +··· for n ∈ N,<br />

S<br />

R<br />

by (i). To finish the proof of (15,) it suffices to show that the last line in (17) is<br />

equal to L o (nP). Letl ∈ Z d . The contribution of l to the last line in (17) equals the<br />

expression<br />

# � {S : l ∈ nS} � − # � {R : l ∈ nR} � +−...<br />

If l �∈ nP, this expression is 0. If l ∈ bd(nP), this expression may be interpreted<br />

(up to the sign) as the Euler characteristic of (a complex of convex polyhedral cones<br />

with apex l <strong>and</strong> union equal to) the polyhedral support cone of P at l. Thus it is 0.<br />

If l ∈ int nP, this expression is (up to the sign) equal to the Euler characteristic of<br />

(a complex of convex polyhedral cones with apex l <strong>and</strong> union equal to) E d . Thus it<br />

is 1. The last line in (17) thus equals L o (nP), concluding the proof of (15) <strong>and</strong> thus<br />

of (ii). ⊓⊔

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

Saved successfully!

Ooh no, something went wrong!