13.11.2014 Views

proofs

proofs

proofs

SHOW MORE
SHOW LESS

Create successful ePaper yourself

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

100 Borsuk’s conjecture<br />

To obtain a general bound for large d, we use monotonicity and unimodality<br />

of the binomial coefficients and the estimates n! > e( n e )n and n! < en( n e )n<br />

(see the appendix to Chapter 2) and derive<br />

∑q−2<br />

( ) ( )<br />

4q − 3 4q<br />

< q<br />

i q<br />

i=0<br />

Thus we conclude<br />

From this, with<br />

f(d) ≥ g(q) =<br />

= q (4q)!<br />

q!(3q)! < q<br />

2 4q−4<br />

q−2 ∑ ( 4q−3<br />

i<br />

i=0<br />

e 4q ( 4q<br />

e ( q<br />

e<br />

e<br />

) 4q<br />

) q<br />

e<br />

( 3q<br />

e<br />

) 3q<br />

= 4q2<br />

e<br />

) > e ( 27<br />

) q.<br />

64q 2 16<br />

( 256<br />

) q.<br />

27<br />

d = (2q − 1)(4q − 3) = 5q 2 + (q − 3)(3q − 1) ≥ 5q 2 for q ≥ 3,<br />

√ √<br />

q = 5 8 + d<br />

8 + 1 64 > d<br />

8 , and ( 27<br />

) √ 1<br />

8<br />

16<br />

> 1.2032,<br />

we get<br />

f(d) ><br />

e<br />

√ √<br />

13d (1.2032) d<br />

> (1.2) d<br />

for all large enough d. □<br />

A counterexample of dimension 560 is obtained by noting that for q = 9 the<br />

quotient g(q) ≈ 758 is much larger than the dimension d(q) = 561. Thus<br />

one gets a counterexample for d = 560 by taking only the “three fourths”<br />

of the points in S that satisfy x 21 + x 31 + x 32 = −1.<br />

Borsuk’s conjecture is known to be true for d ≤ 3, but it has not been<br />

verified for any larger dimension. In contrast to this, it is true up to d = 8<br />

if we restrict ourselves to subsets S ⊆ {1, −1} d , as constructed above<br />

(see [8]). In either case it is quite possible that counterexamples can be<br />

found in reasonably small dimensions.<br />

References<br />

[1] K. BORSUK: Drei Sätze über die n-dimensionale euklidische Sphäre, Fundamenta<br />

Math. 20 (1933), 177-190.<br />

[2] A. HINRICHS & C. RICHTER: New sets with large Borsuk numbers, Discrete<br />

Math. 270 (2003), 137-147.<br />

[3] J. KAHN & G. KALAI: A counterexample to Borsuk’s conjecture, Bulletin<br />

Amer. Math. Soc. 29 (1993), 60-62.<br />

[4] A. NILLI: On Borsuk’s problem, in: “Jerusalem Combinatorics ’93” (H.<br />

Barcelo and G. Kalai, eds.), Contemporary Mathematics 178, Amer. Math.<br />

Soc. 1994, 209-210.<br />

[5] A. M. RAIGORODSKII: On the dimension in Borsuk’s problem, Russian Math.<br />

Surveys (6) 52 (1997), 1324-1325.<br />

[6] O. SCHRAMM: Illuminating sets of constant width, Mathematika 35 (1988),<br />

180-199.<br />

[7] B. WEISSBACH: Sets with large Borsuk number, Beiträge zur Algebra und<br />

Geometrie/Contributions to Algebra and Geometry 41 (2000), 417-423.<br />

[8] G. M. ZIEGLER: Coloring Hamming graphs, optimal binary codes, and the<br />

0/1-Borsuk problem in low dimensions, Lecture Notes in Computer Science<br />

2122, Springer-Verlag 2001, 164-175.

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

Saved successfully!

Ooh no, something went wrong!