30.12.2014 Views

The Maximal Number of Transverse Self-Intersections of Geodesics ...

The Maximal Number of Transverse Self-Intersections of Geodesics ...

The Maximal Number of Transverse Self-Intersections of Geodesics ...

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.

µ 2 3<br />

As an example, we see that a 2 =<br />

3 5<br />

µ 5 8<br />

,a 3 =<br />

8 13<br />

<br />

, all Fibonacci numbers.<br />

Definition 3.4 Let φ = 1+√ 5<br />

, represent the golden ratio.<br />

2<br />

µ <br />

µ <br />

Fn+1<br />

Fn−1<br />

Lemma 3.5 <strong>The</strong> lim n→∞ = φ, lim n→∞ = 1 φ<br />

F n F n<br />

F n−1<br />

+ F n+1<br />

− φ monotonicallydecreaseto2φ − 1 as n →∞.<br />

F n F n<br />

= φ − 1 and the values <strong>of</strong><br />

Lemma 3.6 If W = a n 2 b n 2 , then all cyclic permutations W p1 = a k b n 2 a m and W p2 = b k a n 2 b m<br />

<strong>of</strong> W have symmetric roots.<br />

Pro<strong>of</strong>. We look at equations <strong>of</strong> the form W p1 = a k b n 2 a m and W p2 = b k a n 2 b m . Since n<br />

is even, we know that k + m = n , assuming these equations are cyclic permutations <strong>of</strong> W .<br />

2<br />

<strong>The</strong>n µ we perform the µ following matrix µ multiplication: <br />

F2k−1 F 2k F2n−1 −F 2n F2m−1 F 2m<br />

=<br />

µ<br />

F 2k F 2k+1 −F 2n F 2n+1 F 2m F 2m+1<br />

F2m (F 2k F 2n+1 − F 2n F 2k−1 )+F 2m−1 (−F 2k F 2n + F 2k−1 F 2n−1 )<br />

F 2m (−F 2k F 2n + F 2k+1 F 2n+1 )+F 2m−1 (F 2k F 2n−1 − F 2n F 2k+1 )<br />

<br />

F 2m (−F 2k F 2n + F 2k−1 F 2n−1 )+F 2m+1 (F 2k F 2n+1 − F 2n F 2k−1 )<br />

µ<br />

F 2m (F 2k F 2n−1 −<br />

µ<br />

F 2n F 2k+1 )+F 2m+1<br />

µ<br />

(−F 2k F 2n + F 2k+1 F<br />

2n+1 )<br />

F2k−1 −F 2k F2n−1 F 2n F2m−1 −F 2m<br />

=<br />

µ<br />

−F 2k F 2k+1 F 2n F 2n+1 −F 2m F 2m+1<br />

F2m (F 2k F 2n+1 − F 2n F 2k−1 )+F 2m−1 (−F 2k F 2n + F 2k−1 F 2n−1 )<br />

F 2m (F 2k F 2n − F 2k+1 F 2n+1 )+F 2m−1 (−F 2k F 2n−1 + F 2n F 2k+1 )<br />

<br />

F 2m (F 2k F 2n − F 2k−1 F 2n−1 )+F 2m+1 (−F 2k F 2n+1 + F 2n F 2k−1 )<br />

.<br />

F 2m (F 2k F 2n−1 − F 2n F 2k+1 )+F 2m+1 (−F 2k F 2n + F 2k+1 F 2n+1 )<br />

Since the only changes in the matrices are the signs <strong>of</strong> the upper right and lower left<br />

entries, µ let us assume that the entries are<br />

µ <br />

µ <br />

α β<br />

α β<br />

α −β<br />

, without loss <strong>of</strong> generality. So W<br />

γ δ<br />

p1 = ,andW<br />

γ δ<br />

p2 =<br />

.<br />

−γ δ<br />

By performing our transformation, we seeq<br />

T (W p1 )=z has roots at − (δ − α) ± (δ − α) 2 +4γβ<br />

, while T (W p2 )=z has roots at<br />

q<br />

2γ<br />

− (δ − α) ± (δ − α) 2 +4γβ<br />

. Since the only difference in these two roots is the lower<br />

−2γ<br />

term’s sign, they are symmetric.<br />

Lemma 3.7 Words with n even, n ≥ 4 <strong>of</strong> the form W = a n 2 b n 2<br />

[1, ∞).<br />

have roots in [0, 1] and<br />

12

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

Saved successfully!

Ooh no, something went wrong!