Used and New from Other Sellers
Used and New from Other Sellers
from $19.99
Usually ships in 12 business days
(Save 81%)
Other sellers (Hardcover)

All (6)
from
$19.99

New (2)
from
$19.99

Used (4)
from
$28.75
Note: Marketplace items are not eligible for any BN.com coupons and promotions
Brand new and unread! Join our growing list of satisfied customers!
Ships from: Phoenix, MD
Usually ships in 12 business days
 •Standard, 48 States
 •Standard (AK, HI)
BRAND NEW
Ships from: Avenel, NJ
Usually ships in 12 business days
 •Canadian
 •International
 •Standard, 48 States
 •Standard (AK, HI)
More About This Textbook
Overview
Outer billiards is a basic dynamical system defined relative to a convex shape in the plane. B. H. Neumann introduced this system in the 1950s, and J. Moser popularized it as a toy model for celestial mechanics. All along, the socalled MoserNeumann question has been one of the central problems in the field. This question asks whether or not one can have an outer billiards system with an unbounded orbit. The MoserNeumann question is an idealized version of the question of whether, because of small disturbances in its orbit, the Earth can break out of its orbit and fly away from the Sun. In Outer Billiards on Kites, Richard Schwartz presents his affirmative solution to the MoserNeumann problem. He shows that an outer billiards system can have an unbounded orbit when defined relative to any irrational kite. A kite is a quadrilateral having a diagonal that is a line of bilateral symmetry. The kite is irrational if the other diagonal divides the quadrilateral into two triangles whose areas are not rationally related. In addition to solving the basic problem, Schwartz relates outer billiards on kites to such topics as Diophantine approximation, the modular group, selfsimilar sets, polytope exchange maps, profinite completions of the integers, and solenoidsconnections that together allow for a fairly complete analysis of the dynamical system.
Editorial Reviews
Mathematical Review
This research is an exemplary sample of experimental mathematics.— Serge L. Tabachnikov
Mathematical Review  Serge L. Tabachnikov
This research is an exemplary sample of experimental mathematics.From the Publisher
"This research is an exemplary sample of experimental mathematics."—Serge L. Tabachnikov, Mathematical ReviewProduct Details
Related Subjects
Meet the Author
Richard Evan Schwartz is professor of mathematics at Brown University and the author of "Spherical CR Geometry and Dehn Surgery" (Princeton).
Read an Excerpt
Outer Billiards on Kites
By Richard Evan Schwartz
PRINCETON UNIVERSITY PRESS
Copyright © 2009 Princeton University PressAll right reserved.
ISBN: 9780691142494
Chapter One
Introduction1.1 DEFINITIONS AND HISTORY
B. H. Neumann [N] introduced outer billiards in the late 1950s. In the 1970s, J. Moser [M1] popularized outer billiards as a toy model for celestial mechanics. See [T1], [T3], and [DT1] for expositions of outer billiards and many references on the subject.
Outer billiards is a dynamical system defined (typically) in the Euclidean plane. Unlike the more familiar variant, which is simply called billiards, outer billiards involves a discrete sequence of moves outside a convex shape rather than inside it. To define an outer billiards system, one starts with a bounded convex set K [subset] [R.sup.2] and considers a point [x.sub.0] [member of] [R.sup.2]  K. One defines [x.sub.1] to be the point such that the segment [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] is tangent to K at its midpoint and K lies to the right of the ray [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]. The iteration [x.sub.0] [right arrow] [x.sub.1] [right arrow] [x.sub.2] [right arrow] ... is called the forward outer billiards orbit of [x.sub.0]. It is defined for almost every point of [R.sup.2]  K. The backward orbit is defined similarly.
One important feature of outer billiards is that it is an affinely invariant system. Since affine transformations carry lines to lines and respect the property of bisection, an affine transformation carrying one shape to another conjugates the one outer billiards system to the other.
It is worth recalling here a few basic definitions about orbits. An orbit is called periodic if it eventually repeats itself, and otherwise aperiodic. An orbit is called bounded if the whole orbit lies in a bounded portion of the plane. Otherwise, the orbit is called unbounded. Sometimes (un)bounded orbits are called (un)stable.
J. Moser [M2, p. 11] attributes the following question to Neumann ca. 1960, though it is sometimes called Moser's question. Is there an outer billiards system with an unbounded orbit? This is an idealized version of the question about the stability of the solar system. Here is a chronological list of much of the work related to this question.
J. Moser [M2] sketches a proof, inspired by KAM theory, that outer billiards on K has all bounded orbits provided that [partial derivative]K is at least [ITLITL.sup.6] smooth and positively curved. R. Douady gives a complete proof in his thesis [D].
In VivaldiShaidenko [VS], Kolodziej [Ko], and GutkinSimanyi [GS], it is proved (each with different methods) that outer billiards on a quasirational polygon has all orbits bounded. This class of polygons includes rational polygons  i.e., polygons with rationalcoordinate vertices  and also regular polygons. In the rational case, all defined orbits are periodic.
S. Tabachnikov [T3] analyzes the outer billiards system for a regular pentagon and shows that there are some nonperiodic (but bounded) orbits.
P. Boyland [B] gives examples of [ITLITL.sup.1] smooth convex domains for which an orbit can contain the domain boundary in its [omega]limit set.
F. Dogru and S. Tabachnikov [DT2] show that, for a certain class of polygons in the hyperbolic plane, called large, all outer billiards orbits are unbounded. (One can define outer billiards in the hyperbolic plane, though the dynamics has a somewhat different feel to it.)
D. Genin [G] shows that all orbits are bounded for the outer billiards systems associated to trapezoids. See A.4. Genin also makes a brief numerical study of a particular irrational kite based on the square root of 2, observes possibly unbounded orbits, and indeed conjectures that this is the case.
In [S] we prove that outer billiards on the Penrose kite has unbounded orbits, thereby answering the MoserNeumann question in the affirmative. The Penrose kite is the convex quadrilateral that arises in the Penrose tiling.
Recently, D. Dolgopyat and B. Fayad [DF] showed that outer billiards on a halfdisk has some unbounded orbits. Their proof also works for regions obtained from a disk by nearly cutting it in half with a straight line. This is a second affirmative answer to the MoserNeumann question.
The result in [S] naturally raises questions about generalizations. The purpose of this book is to develop the theory of outer billiards on kites and show that the phenomenon of unbounded orbits for polygonal outer billiards is (at least for kites) quite robust.
1.2 THE ERRATIC ORBITS THEOREM
A kite is a convex quadrilateral K having a diagonal that is a line of symmetry. We say that K is (ir)rational if the other diagonal divides K into two triangles whose areas are (ir)rational multiples of each other. Equivalently, K is rational iff it is affinely equivalent to a quadrilateral with rational vertices. To avoid trivialities, we require that exactly one of the two diagonals of K is a line of symmetry. This means that a rhombus does not count as a kite.
Since outer billiards is an affinely natural system, we find it useful to normalize kites in a particular way. Any kite is affinely equivalent to the quadrilateral K(A) having vertices
(1, 0), (0, 1), (0,1), (A, 0), A [member of] (0, 1). (1.1)
Figure 1.1 shows an example. The omitted case A = 1 corresponds to rhombuses. Henceforth, when we say kite, we mean K(A) for some A. The kite K(A) is (ir)rational iff A is (ir)rational.
Let [Z.sub.odd] denote the set of odd integers. Reflection in each vertex of K(A) preserves R x [Z.sub.odd]. Hence outer billiards on K(A) preserves R x [Z.sub.odd]. We call an outer billiards orbit on K(A) special if (and only if) it is contained in R x [Z.sub.odd]. We discuss only special orbits in this book. The special orbits are hard enough for us already. In the appendix, we will say something about the general case. See A.3.
We call an orbit forward erratic if the forward orbit is unbounded and also returns to every neighborhood of a kite vertex. We state the same definition for the backward direction. We call an orbit erratic if it is both forward and backward erratic. In Parts 14 of the book we will prove the following result.
Theorem 1.1 (Erratic Orbits) The following hold for any irrational kite.
1. There are uncountably many erratic special orbits.
2. Every special orbit is either periodic or unbounded in both directions.
3. The set of periodic special orbits is open dense in R x [Z.sub.odd].
It follows from the work on quasirational polygons cited above that all orbits are periodic relative to a rational kite. (The analysis in this book gives another proof of this fact, at least for special orbits. See the remark at the end of 3.2.) Hence the Erratic Orbits Theorem has the following corollary.
Corollary 1.2 Outer billiards on a kite has an unbounded orbit if and only if the kite is irrational.
The Erratic Orbits Theorem is an intermediate result included so that the reader can learn a substantial theorem without having to read the whole book. We will describe our main result in the next two sections.
1.3 COROLLARIES OF THE COMET THEOREM
In Parts 5 and 6 of the book we will go deeper into the subject and establish our main result, the Comet Theorem. The Comet Theorem and its corollaries considerably sharpen the Erratic Orbits Theorem. We defer statement of the Comet Theorem until the next section. In this section, we describe some of its corollaries.
Given a Cantor set ITLITL contained in a line L, we let [ITLITL.sup.#] be the set obtained from ITLITL by deleting the endpoints of the components of L  C. We call [ITLITL.sup.#] a trimmed Cantor set. Note that ITLITL  [ITLITL.sup.#] is countable.
The interval
I = [0, 2] x {1} (1.2)
turns out to be a very useful interval. Figure 1.2 shows I and its first 3 iterates under the outer billiards map.
Let [U.sub.A] denote the set of unbounded special orbits relative to A.
Theorem 1.3 Relative to any irrational A [member of] (0, 1), the following are true.
1. [U.sub.A] is minimal: Every orbit in [U.sub.A] is dense in [U.sub.A] and all but at most 2 orbits in [U.sub.A] are both forward dense and backward dense in [U.sub.A].
2. [U.sub.A] is locally homogeneous: Every two points in [U.sub.A] have arbitrarily small neighborhoods that are isometric to each other.
3. [U.sub.A] [intersection] I = [ITLITL.sup.#.sub.A] for some Cantor set [ITLITL.sub.A].
Remarks:
(i) One endpoint of [ITLITL.sub.A] is the kite vertex (0,1). Hence Statement 1 implies that all but at most 2 unbounded special orbits are erratic. The remaining special orbits, if any, are each erratic in one direction.
(ii) Statements 2 and 3 combine to say that every point in [U.sub.A] lies in an interval that intersects [U.sub.A] in a trimmed Cantor set. This gives us a good local picture of [U.sub.A]. One thing we are missing is a good global picture of [U.sub.A].
(iii) The Comet Theorem describes [ITLITL.sub.A] explicitly.
Given Theorem 1.3, it makes good sense to speak of the first return map to any interval in R x [Z.sub.odd]. From the minimality result, the local nature of the return map is essentially the same around any point of [U.sub.A]. To give a crisp picture of this first return map, we consider the interval I discussed above.
For j = 1, 2, let [f.sub.j] : [X.sub.j] [right arrow] [X.sub.j] be a map such that [f.sub.j] and [f.sup.1.sub.j] are defined on all but perhaps a finite subset of [X.sub.j]. We call [f.sub.1] and [f.sub.2] essentially conjugate if there are countable sets [ITLITL.sub.j] [subset] [X.sub.j], each one contained in a finite union of orbits, and a homeomorphism
h: [X.sub.1]  [ITLITL.sub.1] [right arrow] [X.sub.2]  [ITLITL.sub.2]
that conjugates [f.sub.1] to [f.sub.2].
An odometer is the map x [right arrow] x + 1 on the inverse limit of the system
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (1.3)
The universal odometer is the map x [right arrow] x + 1 on the profinite completion of Z. This is the inverse limit taken over the system of all finite cyclic groups. For concreteness, Equation 1.3 defines the universal odometer when [D.sub.k] = k factorial. See [H] for a detailed discussion of the universal odometer.
Theorem 1.4 Let [[rho].sub.A] be the first return map to [U.sub.A] [intersection] I.
1. For any irrational A [member of] (0, 1), the map [[rho].sub.A] is defined on all but at most one point and is essentially conjugate to an odometer [Z.sub.A].
2. Any given odometer is essentially conjugate to [[rho].sub.A] for uncountably many difference choices of A.
3. [[rho].sub.A] is essentially conjugate to the universal odometer for almost all A.
Remarks:
(i) The Comet Theorem explicitly describes [Z.sub.A] in terms of a sequence we call the remormalization sequence. This sequence is related to the continued fraction expansion of A. We will give a description of this sequence in the next section. (ii) Theorem 1.4 is part of a larger result. There is a certain suspension flow over the odometer, which we call geodesic flow on the cusped solenoid. It turns out that the timeone map for this flow serves as a good model, in a certain sense, for the dynamics on [U.sub.A]. 24.3.
Our next result highlights an unexpected connection between outer billiards on kites and the modular group [SL.sub.2](Z). The group [SL.sub.2](Z) acts naturally on the upper halfplane model of the hyperbolic plane, [H.sup.2], by linear fractional transformations. Closely related to [SL.sub.2](Z) is the (2, [infinity], [infinity])triangle group [Gamma] generated by reflections in the sides of the geodesic triangle with vertices (0, 1, i). The points 0 and 1 are the cusps, and the point i is the internal vertex corresponding to the right angle of the triangle. See 25.2 for more details. [Gamma] and [SL.sub.2](Z) are commensurable: Their intersection has finite index in both groups. In our next result, we interpret our kite parameter interval (0, 1) as the subset of the ideal boundary of [H.sup.2].
Theorem 1.5 Let S = [0, 1]  Q. Let u(A) be the Hausdorff dimension of [U.sub.A].
1. For all A [member of] S, the set [U.sub.A] has length 0. Hence almost all points in R x [Z.sub.odd] have periodic orbits relative to outer billiards on K(A).
2. If A, A' [member of] S are in the same [Gamma]orbit, then [U.sub.A] and [U.sub.A'] are locally similar. In particular, u(A) = u(A').
3. If A [member of] S is quadratic irrational, then every point of [U.sub.A] lies in an interval that intersects [U.sub.A] in a selfsimilar trimmed Cantor set.
4. The function u is almost everywhere equal to some constant u0 and yet maps every open subset of S onto [0, 1].
Remarks:
(i)We do not know the value of [u.sub.0]. We guess that 0 < [u.sub.0] < 1. Theorem 25.9 gives a formula for u(A) in many cases.
(ii) The word similar in statement 2 means that the two sets have neighborhoods that are related by a similarity. In statement 3, a selfsimilar set is a disjoint finite union of similar copies of itself.
(iii)We will see that statement 2 essentially implies both statements 3 and 4. Statement 2 is the first hint that outer billiards on kites is connected to themodular group. The Comet Theorem says more about this.
(iv) Statement 3 of Theorem 1.4 combines with statement 4 of Theorem 1.5 to say that there is a "typical behavior" for outer billiards on kites, in a certain sense. For almost every parameter A, the dimension of [U.sub.A] is the (unknown) constant [u.sub.0] and the return map [[rho].sub.A] is essentially conjugate to the universal odometer.
We end this section by comparing our results here with the main theorems in [S] concerning the Penrose kite. The Penrose kite parameter is
A = [square root of 5]  2 = [[phi].sup.3],
where [phi] is the golden ratio. In [S], we prove that [ITLITL.sup.#.sub.A] [subset] [U.sub.A] and that the first return map to [ITLITL.sup.#.sub.A] is essentially conjugate to the 2adic odometer. Theorems 1.3 and 1.4 subsume these results about the Penrose kite.
(Continues...)
Table of Contents
Contents
Preface....................xiChapter 1. Introduction....................1
PART 1. THE ERRATIC ORBITS THEOREM....................17
Chapter 2. The Arithmetic Graph....................19
Chapter 3. The Hexagrid Theorem....................33
Chapter 4. Period Copying....................41
Chapter 5. Proof of the Erratic Orbits Theorem....................45
PART 2. THE MASTER PICTURE THEOREM....................53
Chapter 6. The Master Picture Theorem....................55
Chapter 7. The Pinwheel Lemma....................69
Chapter 8. The Torus Lemma....................77
Chapter 9. The Strip Functions....................85
Chapter 10. Proof of the Master Picture Theorem....................93
PART 3. ARITHMETIC GRAPH STRUCTURE THEOREMS....................99
Chapter 11. Proof of the Embedding Theorem....................101
Chapter 12. Extension and Symmetry....................107
Chapter 13. Proof of Hexagrid Theorem I....................117
Chapter 14. The Barrier Theorem....................125
Chapter 15. Proof of Hexagrid Theorem II....................133
Chapter 16. Proof of the Intersection Lemma....................143
PART 4. PERIODCOPYING THEOREMS....................151
Chapter 17. Diophantine Approximation....................153
Chapter 18. The Diophantine Lemma....................163
Chapter 19. The Decomposition Theorem....................171
Chapter 20. Existence of Strong Sequences....................181
PART 5. THE COMET THEOREM....................185
Chapter 21. Structure of the Inferior and Superior Sequences....................187
Chapter 22. The Fundamental Orbit....................193
Chapter 23. The Comet Theorem....................205
Chapter 24. Dynamical Consequences....................219
Chapter 25. Geometric Consequences....................227
PART 6. MORE STRUCTURE THEOREMS....................237
Chapter 26. Proof of the Copy Theorem....................239
Chapter 27. Pivot Arcs in the Even Case....................249
Chapter 28. Proof of the Pivot Theorem....................259
Chapter 29. Proof of the Period Theorem....................273
Chapter 30. Hovering Components....................279
Chapter 31. Proof of the Low Vertex Theorem....................287
Appendix....................295
Bibliography....................303
Index....................305