Sie sind auf Seite 1von 408

Fractals in Probability and Analysis

CHRISTOPHER J. BISHOP

Stony Brook University, Stony Brook, NY

YUVAL PERES

Microsoft Research, Redmond, WA

and Analysis CHRISTOPHER J. BISHOP Stony Brook University, Stony Brook, NY YUVAL PERES Microsoft Research, Redmond,
University Printing House, Cambridge CB2 8BS, United Kingdom One Liberty Plaza, 20th Floor, New York,

University Printing House, Cambridge CB2 8BS, United Kingdom One Liberty Plaza, 20th Floor, New York, NY 10006, USA 477 Williamstown Road, Port Melbourne, VIC 3207, Australia 4843/24, 2nd Floor, Ansari Road, Daryaganj, Delhi - 110002, India 79 Anson Road, #06-04/06, Singapore 079906

Cambridge University Press is part of the University of Cambridge.

It furthers the University’s mission by disseminating knowledge in the pursuit of education, learning, and research at the highest international levels of excellence.

www.cambridge.org Information on this title: www.cambridge.org/9781107134119

10.1017/9781316460238

c Christopher J. Bishop and Yuval Peres 2017

This publication is in copyright. Subject to statutory exception and to the provisions of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press.

First published 2017

Printed in the United Kingdom by Clays, St Ives plc

A catalogue record for this publication is available from the British Library .

ISBN 978-1-107-13411-9 Hardback

Cambridge University Press has no responsibility for the persistence or accuracy of URLs for external or third-party Internet Web sites referred to in this publication and does not guarantee that any content on such Web sites is, or will remain, accurate or appropriate.

Contents

Preface

page

1 Minkowski and Hausdorff dimensions

1

1.1 Minkowski dimension

1

1.2 Hausdorff dimension and the Mass Distribution Principle

4

1.3 Sets defined by digit restrictions

9

1.4 Billingsley’s Lemma and the dimension of measures

17

1.5 Sets defined by digit frequency

21

1.6 Slices

26

1.7 Intersecting translates of Cantor sets *

29

1.8 Notes

34

1.9 Exercises

36

2 Self-similarity and packing dimension

45

2.1 Self-similar sets

45

2.2 The open set condition is sufficient

51

2.3 Homogeneous sets

54

2.4 Microsets

57

2.5 Poincar´e sets *

61

2.6 Alternative definitions of Minkowski dimension

67

2.7 Packing measures and dimension

71

2.8 When do packing and Minkowski dimension agree?

74

2.9 Notes

76

2.10 Exercises

78

3 Frostman’s theory and capacity

83

3.1 Frostman’s Lemma

83

3.2 The dimension of product sets

88

3.3 Generalized Marstrand Slicing Theorem

90

viii

Contents

3.5 Marstrand’s Projection Theorem

95

3.6 Mapping a tree to Euclidean space preserves capacity

100

3.7 Dimension of random Cantor sets

104

3.8 Notes

112

3.9 Exercises

115

4

Self-affine sets

119

4.1 Construction and Minkowski dimension

119

4.2 The Hausdorff dimension of self-affine sets

121

4.3 A dichotomy for Hausdorff measure

125

4.4 The Hausdorff measure is infinite *

127

4.5 Notes

131

4.6 Exercises

133

5

Graphs of continuous functions

136

5.1 Holder¨

continuous functions

136

5.2 The Weierstrass function is nowhere differentiable

140

5.3 estimates

Lower Holder¨

145

5.4 Notes

149

5.5 Exercises

151

6

Brownian motion, Part I

160

6.1 Gaussian random variables

160

6.2 L´evy’s construction of Brownian motion

163

6.3 Basic properties of Brownian motion

167

6.4 Hausdorff dimension of the Brownian path and graph

172

6.5 Nowhere differentiability is prevalent

176

6.6 Strong Markov property and the reflection principle

178

6.7 Local extrema of Brownian motion

180

6.8 Area of planar Brownian motion

181

6.9 General Markov processes

183

6.10 Zeros of Brownian motion

185

6.11 Harris’ inequality and its consequences

189

6.12 Points of increase

192

6.13 Notes

196

6.14 Exercises

199

7

Brownian motion, Part II

201

7.1 Dimension doubling

201

7.2 The Law of the Iterated Logarithm

206

7.3 Skorokhod’s Representation

209

7.4 Donsker’s Invariance Principle

216

 

Contents

ix

7.6 The maximum principle for harmonic functions

226

7.7 The Dirichlet problem

227

7.8 Polar points and recurrence

228

7.9 Conformal invariance *

230

7.10 Capacity and harmonic functions

235

7.11 Notes

239

7.12 Exercises

241

8 Random walks, Markov chains and capacity

244

8.1 Frostman’s theory for discrete sets

244

8.2 Markov chains and capacity

250

8.3 Intersection equivalence and return times

254

8.4 Lyons’ Theorem on percolation on trees

258

8.5 Dimension of random Cantor sets (again)

260

8.6 Brownian motion and Martin capacity

264

8.7 Notes

266

8.8 Exercises

266

9 Besicovitch–Kakeya sets

270

9.1 Existence and dimension

270

9.2 Splitting triangles

276

9.3 Fefferman’s Disk Multiplier Theorem *

278

9.4 Random Besicovitch sets

286

9.5 Projections of self-similar Cantor sets

290

9.6 The open set condition is necessary *

297

9.7 Notes

302

9.8 Exercises

305

10 The Traveling Salesman Theorem

313

10.1 Lines and length

313

10.2 The β -numbers

318

10.3 Counting with dyadic squares

322

10.4 β and µ are equivalent

325

10.5 β -sums estimate minimal paths

329

10.6 Notes

334

10.7 Exercises

337

Appendix A

Banach’s Fixed-Point Theorem

343

Appendix B

Frostman’s Lemma for analytic sets

353

Appendix C

Hints and solutions to selected exercises

360

References

379

Preface

The aim of this book is to acquaint readers with some fractal sets that arise naturally in probability and analysis, and the methods used to study them. The book is based on courses taught by the authors at Yale, Stony Brook University, the Hebrew University and UC Berkeley. We owe a great debt to our advisors, Peter Jones and Hillel Furstenberg; thus the book conveys some of their per- spectives on the subject, as well as our own. We have made an effort to keep the book self-contained. The only prerequi- site is familiarity with measure theory and probability at the level acquired in a first graduate course. The book contains many exercises of varying difficulty. We have indicated with a “” those for which a solution, or a hint, is given in Appendix C. A few sections are technically challenging and not needed for subsequent sections, so could be skipped in the presentation of a given chapter. We mark these with a “*” in the section title.

Acknowledgments: We are very grateful to Ton´ci Antunovi´c, Subhroshekhar Ghosh and Liat Kessler for helpful comments and crucial editorial work. We also thank Ilgar Eroglu, Hrant Hakobyan, Michael Hochman, Nina Holden, Pertti Mattila, Elchanan Mossel, Boris Solomyak, Perla Sousi, Ryokichi Tan- aka, Tatiana Toro, B´alint Vir´ag, Samuel S. Watson, Yimin Xiao and Alex Zhai for useful comments. Rich´ard Balka carefully read the entire manuscript and provided hundreds of detailed corrections and suggestions. Many thanks to David Tranah and Sam Harrison at Cambridge University Press for numerous helpful suggestions. Finally, we dedicate this book to our families: Cheryl, David and Emily Bishop, and Deborah, Alon and Noam Peres; without their support and under- standing, it would have taken even longer to write.

1

Minkowski and Hausdorff dimensions

In this chapter we will define the Minkowski and Hausdorff dimensions of a set

and will compute each in a few basic examples. We will then prove Billings- ley’s Lemma and the Law of Large Numbers. These allow us to deal with more sophisticated examples: sets defined in terms of digit frequencies, ran- dom slices of the Sierpi´nski gasket, and intersections of random translates of the middle thirds Cantor set with itself. Both Minkowski and Hausdorff di- mensions measure how efficiently a set K can be covered by balls. Minkowski dimension requires that the covering be by balls all of the same radius. This

makes it easy to compute, but it lacks certain desirable properties. In the def- inition of Hausdorff dimension we will allow coverings by balls of different radii. This gives a better behaved notion of dimension, but (as we shall see) it

is

usually more difficult to compute.

1.1 Minkowski dimension

A

be

is

denote the minimal number of sets of diameter at most ε needed to cover K . We define the upper Minkowski dimension as

the same as being a bounded set. For a totally bounded set K , let N (K , ε)

covered by a finite number of balls of diameter ε. For Euclidean space, this

subset K of a metric space is called totally bounded if for any ε > 0, it can

dim M (K ) = limsup

ε0

and the lower Minkowski dimension

dim M (K ) = liminf

ε0

1

log N (K , ε)

log1/ε

,

log N (K , ε)

log1/ε

.

05 Feb 2017 at 20:23:35,

02

( K ) = liminf ε → 0 1 log N ( K , ε )

2

Minkowski and Hausdorff dimensions

If the two values agree, the common value is simply called the Minkowski dimension of K and denoted by dim M (K ). When the Minkowski dimension of a set K exists, the number of sets of diameter ε needed to cover K grows

like ε dim M ( K )+ o(1)

as ε 0.

We get the same values of dim M (K ) and dim M (K ) if we replace N (K , ε)

by N B (K , ε), which is the number of closed balls of radius ε needed to cover

K . This is because N B (K , ε) N (K , ε) N (K , ε/2)

ball of at most twice the diameter and any ball of radius ε/2 has diameter at most ε; strict inequality could hold in a metric space). For subsets of Euclidean space we can also count the number of axis-parallel squares of side length ε needed to cover K , or the number of such squares taken from a grid. Both pos- sibilities give the same values for upper and lower Minkowski dimension, and for this reason Minkowski dimension is sometimes called the box counting dimension . It is also easy to see that a bounded set A and its closure A satisfy dim M (A ) = dim M (A ) and dim M (A ) = dim M (A). If X is a set and x , y X implies |x y | ≥ ε, we say X is ε -separated. Let N sep (K , ε) be the number of elements in a maximal ε-separated subset X of K . Clearly, any set of diameter ε/2 can contain at most one point of an ε- separated set X , so N sep (K , ε) N (K , ε/2). On the other hand, every point of K is within ε of a maximal ε-separated subset X (otherwise add that point to X ). Thus N (K , ε) N sep (K , ε). Therefore replacing N (K , ε) by N sep (K , ε) in the definition of upper and lower Minkowski dimension gives the same values (and it is often easier to give a lower bound in terms of separated sets).

(any set is contained in a

Example 1.1.1 Suppose that K is a finite set. Then N (K , ε) is bounded and dim M (K ) exists and equals 0.

Suppose K = [0, 1]. Then at least 1/ε intervals of length ε are

needed to cover K and clearly ε 1 +1 suffice. Thus dim M (K ) exists and equals

Example 1.1.2

1. Similarly, any bounded set in R d with interior has Minkowski dimension d .

Example 1.1.3 Let C be the usual middle thirds Cantor set obtained as fol-

lows. Let C 0 = [0, 1] and define C 1 = [0, 3 ] [ 2 ,1] C 0 by removing the

central interval of length

3 1 . In general, C n is a union of 2 n intervals of length

3 n and C n +1 is obtained by removing the central third of each. This gives a decreasing nested sequence of compact sets whose intersection is the desired set C. The construction gives a covering of C that uses 2 n intervals of length 3 n . Thus for 3 n ε < 3 n+1 we have

1

3

N (C ,ε) 2 n ,

05 Feb 2017 at 20:23:35,

02

3 − n ≤ ε < 3 − n + 1 we have 1 3 N

1.1 Minkowski dimension

3

1.1 Minkowski dimension 3 Figure 1.1.1 The Cantor middle thirds construction. and hence dim M (

Figure 1.1.1 The Cantor middle thirds construction.

and hence

dim M (C ) log2 log3 .

Conversely, the centers of the n th generation intervals form a 3 n -separated set

of size 2 n , so N sep (C ,3 n ) 2 n . Thus

dim M (C ) log2 log3 = log 3 2.

Therefore the Minkowski dimension exists and equals this common value. If

at each stage we remove the middle α (0 < α < 1) we get a Cantor set C α with

Minkowski dimension log2/(log2 + log

1

1α ).

Example 1.1.4 Consider K = {0}∪{1,

1

2 ,

1

1

3 , 4 ,

}. Observe that

1

1

1

1

n 1 n = n (n 1) >

n 2 .

So, for ε > 0, if we choose n so that (n + 1) 2 < ε n 2 , then n ε 1/2 and

n distinct intervals of length ε are needed to cover the points 1,

n . The

interval [0, n +1 ] can be covered by n + 1 additional intervals of length ε. Thus

1

2 ,

1

,

1

ε 1/2 N (K ,ε) 2ε 1/2 + 1.

Hence dim M (K ) = 1/2.

This example illustrates a drawback of Minkowski dimension: finite sets have dimension zero, but countable sets can have positive dimension. In par- ticular, it is not true that dim M ( n E n ) = sup n dim M (E n ), a useful property for a dimension to have. In the next section, we will introduce Hausdorff di- mension, which does have this property (Exercise 1.6). In the next chapter, we will introduce packing dimension, which is a version of upper Minkowski dimension forced to have this property.

05 Feb 2017 at 20:23:35,

02

dimension, which is a version of upper Minkowski dimension forced to have this property. 05 Feb

4

Minkowski and Hausdorff dimensions

1.2 Hausdorff dimension and the Mass Distribution Principle

Given any set K in a metric space, we define the α -dimensional Hausdorff content as

α (K ) = inf |U i | α : K

H

i

i

U i ,

where {U i } is a countable cover of K by any sets and |E | denotes the diameter of a set E .

Definition 1.2.1

The Hausdorff dimension of K is defined to be

dim(K ) = inf{α : H

α (K ) = 0}.

More generally we define

α (K ) = inf |U i | α : K U i ,|U i | < ε ,

H

ε

i

i

where each U i is now required to have diameter less than ε. The α -dimensional Hausdorff measure of K is defined as

H α (K ) = lim

ε0 H

ε

α

(K ).

This is an outer measure; an outer measure on a non-empty set X is a function μ from the family of subsets of X to [0, ] that satisfies

μ (0/ ) = 0,

μ

μ ( =1 A j )

(A ) μ (B ) if A B ,

j

j =1

μ (A j ).

For background on real analysis see Folland (1999). The α-dimensional Haus- dorff measure is even a Borel measure in R d ; see Theorem 1.2.4 below. When α = d N, then H α is a constant multiple of L d , d -dimensional Lebesgue measure.

α (K )

If we admit only open sets in the covers of K , then the value of H

ε

does not change. This is also true if we only use closed sets or only use convex

sets. Using only balls might increase H

α by at most a factor of 2 α , since any

ε

set K is contained in a ball of at most twice the diameter. Still, the values for

which H α (K ) = 0 are the same whether we allow covers by arbitrary sets or only covers by balls.

Definition 1.2.2 Let μ be an outer measure on X . A set K in X is μ - measurable , if for Every set A X we have

H α (A ) = μ (A K )+ μ (A K c ).

05 Feb 2017 at 20:23:35,

02

X we have H α ( A ) = μ ∗ ( A ∩ K )+

1.2

Hausdorff dimension and the Mass Distribution Principle

5

Definition 1.2.3 Let (Ω, d ) be a metric space. An outer measure μ on Ω is called a metric outer measure if dist(A , B ) > 0 =μ(A B) = μ(A )+ μ(B ), where A and B are two subsets of Ω.

Since Hausdorff measure H α is clearly a metric outer measure, the follow- ing theorem shows that all Borel sets are H α -measurable. This implies that H α is a Borel measure (see Folland (1999)).

Theorem 1.2.4 Let μ be a metric outer measure. Then all Borel sets are μ -measurable.

Proof It suffices to show any closed set K is μ-measurable, since the measur- able sets form a σ -algebra. So, let K be a closed set. We must show for any set A Ω with μ(A ) < ,

μ(A ) μ(A K )+ μ(A K c ).

 

(1.2.1)

Let B 0 = 0/

and for n 1 define B n = { x A : dist(x , K ) >

1

n }, so that

 

 

B n = A K c

n

=1

(since K is closed). Since μ is a metric outer measure and B n A \ K ,

 

μ(A ) μ[(A K ) B n ] = μ(A K )+ μ(B n ).

(1.2.2)

For all m N, the sets D n = B n \ B n1 satisfy

m

=1 μ(D 2 j ) = μ

j

=1 D 2 j μ(A ),m,

m

j

since if x B n , and y D n +2 , then

dist(x , y ) dist(x , K ) dist(y, K ) 1 n

1

n + 1 .

Similarly m =1 μ(D 2 j 1 ) μ(A ). So =1 μ(D j ) < . The inequality

j

j

μ(B n ) μ(A K c ) μ(B n )+

j = n +1

μ(D j )

implies that μ(B n ) μ(A K c ) as n . Thus letting n in (1.2.2) gives

(1.2.1).

μ ( A ∩ K c ) as n → ∞ . Thus letting n →

05 Feb 2017 at 20:23:35,

02

μ ( A ∩ K c ) as n → ∞ . Thus letting n →

6

Minkowski and Hausdorff dimensions

The construction of Hausdorff measure can be made a little more general by considering a positive, increasing function ϕ on [0, ) with ϕ(0) = 0. This is called a gauge function and we may associate to it the Hausdorff content

ϕ

H

(K ) = inf

i

ϕ(|U i |) : K

i

i ;

U

then H

ϕ (K ) are defined as before. The case

ϕ(t ) = t α is just the case considered above. We will not use other gauge func- tions in the first few chapters, but they are important in many applications, e.g., see Exercise 1.59 and the Notes for Chapter 6.

ϕ (K ), and H ϕ (K ) = lim ε0 H

ε

ε

Lemma 1.2.5

If H α (K ) < then H β (K ) = 0 for any β > α .

Proof

α

It follows from the definition of H

ε

that

β

H

ε

(K ) ε βα H

α

ε

(K

),

which gives the desired result as ε 0.

α ε ( K ) , which gives the desired result as ε → 0. Thus

Thus if we think of H α (K ) as a function of α, the graph of H α (K ) versus α shows that there is a critical value of α where H α (K ) jumps from to 0. This critical value is equal to the Hausdorff dimension of the set. More generally we have:

Proposition 1.2.6 For every metric space E we have

and therefore

H α (E ) = 0

α

H

(E ) = 0

dim E = inf{α : H α (E ) = 0} = inf{α : H α (E ) < }

= sup{α : H α (E ) > 0} = sup{α : H α (E ) = }.

(E ) = 0,

Proof Since H α (E )

then for every δ > 0 there is a covering of E by sets { E k } with =1 |E k | α < δ. These sets have diameter less than δ 1/α , hence H 1/α (E ) < δ . Letting δ 0

α

H

(E ), it suffices to prove “”. If H

α

α

δ

k

yields H α (E ) = 0, proving the claimed equivalence. The equivalence readily implies that dim E = inf{α : H α (E ) = 0} = sup{α : H α (E ) > 0}. The other conclusions follow from Lemma 1.2.5.

> 0 } . The other conclusions follow from Lemma 1.2.5. The following relationship to Minkowski

The following relationship to Minkowski dimension is clear

dim(K ) dim M (K ) dim M (K ).

05 Feb 2017 at 20:23:35,

02

(1.2.3)

dimension is clear dim ( K ) ≤ dim M ( K ) ≤ dim M

1.2

Hausdorff dimension and the Mass Distribution Principle

7

Indeed, if B i = B (x i , ε/2) are N (K , ε) balls of radius ε/2 and centers x i in K that cover K , then consider the sum

S ε =

N ( K ,ε)

| B i | α = N (K ,ε)ε α = ε α R ε ,

i=1

where R ε = log N ( K ,ε)

log(1/ε)

. If α > liminf ε0 R ε = dim M (K ) then inf ε>0 S ε = 0.

Strict inequalities in (1.2.3) are possible.

Example 1.2.7 Example 1.1.4 showed that K = {0} n { 1 n } has Minkowski

dimension 2 . However, any countable set has Hausdorff dimension 0, for if we

} and cover the n th point by a ball of diameter

as small as we wish for any α > 0. Thus K is

a compact set for which the Minkowski dimension exists, but is different from

the Hausdorff dimension.

δ n = ε2 n we can make n δ α

enumerate the points { x 1 , x 2 ,

1

n

Lemma 1.2.8 (Mass Distribution Principle) If E supports a strictly positive Borel measure μ that satisfies

μ(B (x , r )) Cr α ,

for some constant 0 < C < and for every ball B (x , r ), then H α (E ) H

α (E )

μ(E )/C. In particular, dim(E )

α .

Let {U i } be a cover of E . For { r i }, where r i > |U i |, we look at the

Proof

following cover: choose x i in each U i , and take open balls B(x i , r i ). By as-

sumption,

μ(U i ) μ(B (x i , r i )) Cr i α .

We deduce that μ (U i ) C |U i | α , whence

i

|U i | α

i

Thus H α (E ) H

α (E ) μ(E )/C .

μ(U i )

C

μ(E )

C

.

) ≥ μ ( E ) / C . μ ( U i ) C ≥

We note that upper bounds for Hausdorff dimension usually come from find- ing explicit coverings of the set, but lower bounds are proven by constructing an appropriate measure supported on the set. Later in this chapter we will gen- eralize the Mass Distribution Principle by proving Billingsley’s Lemma (Theo- rem 1.4.1) and will generalize it even further in later chapters. As a special case of the Mass Distribution Principle, note that if A R d has positive Lebesgue d -measure then dim(A ) = d .

05 Feb 2017 at 20:23:35,

02

note that if A ⊆ R d has positive Lebesgue d -measure then dim ( A

8

Minkowski and Hausdorff dimensions

Example 1.2.9 Consider the Cantor set E obtained by replacing the unit square in the plane by four congruent sub-squares of side length α < 1/2 and continuing similarly. See Figure 1.2.1. We can cover the set by 4 n squares of diameter 2·α n . Thus

dim M (E ) lim

n

log4 n

log4

log( 2α n ) =

logα .

On the other hand, it is also easy to check that at least 4 n sets of diameter α n are needed, so

dim M (E )

log4

logα .

Thus the Minkowski dimension of this set equals β = log4/ logα.

dimension of this set equals β = − log4 / log α . Figure 1.2.1 Four

Figure 1.2.1 Four generations of a Cantor set.

We automatically get dim(E ) β and we will prove the equality using Lemma 1.2.8. Let μ be the probability measure defined on E that gives each nth generation square the same mass (namely 4 n ). We claim that

μ(B (x , r )) Cr β ,

for all disks and some 0 < C < . To prove this, suppose B = B (x , r ) is some disk hitting E and choose n so that α n +1 r < α n . Then B can hit at most 4 of the n th generation squares and so, since α β = 1/4,

μ(B E ) 4·4 n = 4α nβ 16 r β .

Example 1.2.10 Another simple set for which the two dimensions agree and are easy to compute is the von Koch snowflake. To construct this we start with an equilateral triangle. At each stage we add to each edge an equilateral triangle pointing outward of side length 1/3 the size of the current edges and centered on the edge. See Figure 1.2.2 for the first four iterations of this process. The boundary of this region is a curve with dimension log4/ log3 (see Theorem 2.2.2). We can also think of this as a replacement construction, in which at

05 Feb 2017 at 20:23:35,

02

log3 (see Theorem 2.2.2). We can also think of this as a replacement construction, in which

1.3 Sets defined by digit restrictions

9

each stage, a line segment is replaced by an appropriately scaled copy of a polygonal curve.

by an appropriately scaled copy of a polygonal curve. Figure 1.2.2 Four generations of the von
by an appropriately scaled copy of a polygonal curve. Figure 1.2.2 Four generations of the von
by an appropriately scaled copy of a polygonal curve. Figure 1.2.2 Four generations of the von
by an appropriately scaled copy of a polygonal curve. Figure 1.2.2 Four generations of the von

Figure 1.2.2 Four generations of the von Koch snowflake.

Even for some relatively simple sets the Hausdorff dimension is still un- known. Consider the Weierstrass function (Figure 1.2.3)

f α,b (x ) =

n=1

b n α cos(b n x),

where b > 1 is real and 0 < α < 1. It is conjectured that the Hausdorff dimen- sion of its graph is 2 α, and this has been proven when b is an integer; see the discussion in Example 5.1.7. On the other hand, some sets that are more difficult to define, such as the graph of Brownian motion (Figure 1.2.4), will turn out to have easier dimensions to compute (3/2 by Theorem 6.4.3).

have easier dimensions to compute (3 / 2 by Theorem 6.4.3). Figure 1.2.3 The Weierstrass function

Figure 1.2.3 The Weierstrass function with b = 2, α = 1/2. This graph has Minkowski dimension 3/2 and is conjectured to have the same Hausdorff dimen- sion.

1.3 Sets defined by digit restrictions

In this section we will consider some more complicated sets for which the Minkowski dimension is easy to compute, but the Hausdorff dimension is not

05 Feb 2017 at 20:23:35,

02

sets for which the Minkowski dimension is easy to compute, but the Hausdorff dimension is not

10

Minkowski and Hausdorff dimensions

10 Minkowski and Hausdorff dimensions Figure 1.2.4 1-dimensional Brownian motion. This graph has dimension 3/2 al-

Figure 1.2.4 1-dimensional Brownian motion. This graph has dimension 3/2 al- most surely.

so obvious, and will be left to later sections. These subsets of [0, 1] will be defined by restricting which digits can occur at a certain position of a num- ber’s b -ary expansion. In a later section we will consider sets defined by the asymptotic distribution of the digits. We start by adapting Hausdorff measures to b -adic grids. Let b 2 be an integer and consider b -adic expansions of real numbers, i.e.,

to each sequence { x n }∈{0,1,

b 1} N we associate the real number

,

x n b n [0,1].

x =

n =1

b-adic expansions give rise to Cantor sets by restricting the digits we are al- lowed to use. For example, if we set b = 3 and require x n ∈ {0, 2} for all n we get the middle thirds Cantor set C .

For each integer n let I n (x ) denote the unique half-open interval of the form

n ) containing x . Such intervals are called b -adic intervals of generation

b n , n ( dyadic if b = 2). It has been observed (by Frostman (1935) and Besicovitch (1952)) that we can restrict the infimum in the definition of Hausdorff measure to coverings of the set that involve only b -adic intervals and only change the value by a bounded factor. The advantage of dealing with these intervals is that they are nested, i.e., two such intervals either are disjoint or one is contained in the other. In particular, any covering by b-adic intervals always contains a subcover by disjoint intervals (just take the maximal intervals). Furthermore, the b -adic intervals can be given the structure of a tree, an observation that we will use extensively in later chapters.

[ k 1

k

b

05 Feb 2017 at 20:23:35,

02

of a tree, an observation that we will use extensively in later chapters. [ k −

1.3 Sets defined by digit restrictions

We define the grid Hausdorff content by

˜

H

α

(A ) = inf{ |J i | α , A J i },

i i

and the grid Hausdorff measures by

˜

H

α

ε

(A ) = inf{ | J i | α , A J i ,|J i | < ε},

i

H α (A ) = lim H

ε0

α

ε

˜

˜

i

(A

),

11

(1.3.1)

(1.3.2)

(1.3.3)

where the infimums are over all coverings of A R by collections { J i } of b - adic intervals. The grid measure depends on b , but we omit it from the notation; usually the value of b is clear from context. Clearly

˜

H

α (A ) H α (A ),

since we are taking the infimum over a smaller set of coverings. However, the two sides are almost of the same size, i.e.,

˜

H

α (A ) (b + 1)H α (A ),

since any interval I can be covered by at most (b + 1) shorter b-adic intervals;

just take the smallest n so that b n ≤ | I | and take all b -adic intervals of length

˜

b n that hit I . Thus H α (A ) = 0 if and only if H α (A ) = 0. Similarly, we can

define N (K , ε) to be the minimal number of closed b-adic intervals of length ε

needed to cover K . We get

N (K ,ε) N (K ,ε) (b + 1)N (K ,ε).

Hence, the Hausdorff and the Minkowski dimensions are not changed.

(A ) and grid Hausdorff measures

H

ϕ (A ) and H ϕ (A ) with respect to any gauge function ϕ by replacing | J i | α

with ϕ(| J i |) in (1.3.1) and (1.3.2). Furthermore, the definitions can be extended to subsets of R n by replacing the b -adic intervals J i in (1.3.1) and (1.3.2) with

b -adic cubes (products of b -adic intervals of equal length).

Definition 1.3.1

˜

˜

One can define grid Hausdorff content H

˜

˜

ε

˜

ϕ

For S N the upper density of S is

N })

#(S ∩{1,

d(S ) = limsup

N

,

N

.

Here and later #(E ) denotes the number of elements in E , i.e., # is counting measure. The lower density is

d(S ) = liminf

N

#(S ∩{1,

,

N })

N

.

05 Feb 2017 at 20:23:35,

02

measure. The lower density is d ( S ) = liminf N → ∞ # (

12

Minkowski and Hausdorff dimensions

If d(S ) = d(S ), then the limit exists and is called d (S ), the density of S .

Example 1.3.2

Suppose S N, and define

A S = x = S x k 2 k : x k ∈ {0,1} .

k

The set A S is covered by exactly

2

n

k

=1 1 S ( k ) = 2 #( S ∩{1,

,n

})

closed dyadic intervals of generation n , where 1 S is the characteristic function

of S , i.e., 1 S (n ) = 1 for n S , and 1 S (n ) = 0 for n

S . So

Thus

which implies

˜

log 2 N (A S ,2 n ) =

n

1 S (k ).

˜

log N (A S ,2 n )

= 1

k

log2

n

n

k

 

dim

M

(A S ) =

dim

M

=1

n

1 S (k ),

=1

d(S ),

(A S ) = d(S ).

It is easy to construct sets S where the liminf and limsup differ and hence we get compact sets where the Minkowski dimension does not exist (see Exercise 1.17). The Hausdorff dimension of A S is equal to the lower Minkowski dimen- sion. We shall prove this in the next section as an application of Billingsley’s Lemma.

We can also construct A S as follows. Start with the interval [0, 1] and sub- divide it into two equal length subintervals [0, 1/2] and [1/2, 1]. If 1 S then

S then keep only the leftmost, [0, 1/2]. Cut each

keep both intervals and if 1

of the remaining intervals in half, keeping both subintervals if 2 S and only keeping the left interval otherwise. In general, at the n th step we have a set A S A S 1 that is a finite union of intervals of length 2 n (some may be adjoin- ing). We cut each of the intervals in half, keeping both subintervals if n S and

n

n

throwing away the right-hand one if n

n . In Figure 1.3.1 we have drawn the first generations of the construction corre-

sponding to S = {1, 3, 4, 6, 8, 10,

S . The limiting set is A S =

n

=1 A S

}.

Example 1.3.3

The shifts of finite type are defined by restricting which

05 Feb 2017 at 20:23:35,

02

n = 1 A S } . Example 1.3.3 The shifts of finite type are defined

1.3 Sets defined by digit restrictions

13

1.3 Sets defined by digit restrictions 13 Figure 1.3.1 First 8 generations of A S for

Figure 1.3.1 First 8 generations of A S for S = {1, 3, 4, 6, 8, 10,

}.

digits can follow other digits. Let A = (A i j ), 0 i, j < b , be a b × b matrix of 0s and 1s, and define

X A =

n=1 x n b n : A x n x n+1 = 1 for all n 1 .

We will also assume that if the j th column of A is all zeros, so is the j th row; this implies every finite sequence { x n } n satisfying the condition above can be extended to at least one infinite sequence satisfying the condition. Thus such finite strings correspond to b -adic closed dyadic intervals that intersect X A .

Notice that a b -ary rational number r will belong to X A if either of its two possible b -ary expansions does (one is eventually all 0s, the other eventually all (b 1)s). The condition on { x n } described in the example is clearly shift invariant and is the intersection of countably many “closed” conditions, so the set X A is compact and invariant under the map T b , where T b (x )=(bx ) mod 1. For example, if

1

A = 1

1

0 ,

1

then x X A if and only if the binary representation of x has no two consecutive 1s. The first ten generations in the construction of the Cantor set X A are shown in Figure 1.3.3. Any such matrix can be represented by a directed graph where

the vertices represent the numbers {0,

, i to j if A i j = 1. For example, the 2 × 2 matrix A above is represented by

Figure 1.3.2. An element of x = n x n b n X A corresponds in a natural way to an infinite

path { x 1 , x 2 ,

} on this graph and a b -adic interval hitting X A corresponds to

a finite path. Conversely, any finite path of length n corresponds to a b -adic

b 1} and a directed edge connects

˜

interval of length b n hitting X A . Thus N n (X A ) N (X A , b n ) is the number of

05 Feb 2017 at 20:23:35,

02

X A . Thus N n ( X A ) ≡ N ( X A ,

14

Minkowski and Hausdorff dimensions

0 1
0
1

Figure 1.3.2 The directed graph representing A .

0 1 Figure 1.3.2 The directed graph representing A . Figure 1.3.3 Approximations of X A

Figure 1.3.3 Approximations of X A .

distinct paths of length n in the graph (this uses our assumption about the rows and columns of A ). By the definition of matrix multiplication, the number of paths of length n from i to j is (A n ) i j (i.e., the (ith, j th) entry in the matrix A n ). (See Exercise 1.23.) Thus the total number of length n paths is N n (X A ) = A n , where the norm of a matrix is defined as B = i, j |B i j |. (However, since any two norms on a finite-dimensional space are comparable, the precise norm won’t matter.) Thus

where

dim M (X A ) = lim

n

log N n (X A ) n log b

= lim

n

log(N n (X A )) 1/ n log b

= logρ(A)

log b

ρ(A ) = lim

n

A n 1/ n

,

(1.3.4)

is, by definition, the spectral radius of the matrix A , and is equal to the ab- solute value of the largest eigenvalue. That this limit exists is a standard fact about matrices and is left to the reader (Exercise 1.24). For the matrix

A = 1

1

0 ,

1

the spectral radius is (1 + 5)/2 (Exercise 1.25), so the Minkowski dimension

05 Feb 2017 at 20:23:35,

02

, 1 the spectral radius is ( 1 + √ 5 ) / 2 (Exercise 1.25),

1.3 Sets defined by digit restrictions

of X A is

logρ(A )

log(1 + 5) log2

=

log2

log2

.

15

We shall see in Example 2.3.3, that for shifts of finite type the Hausdorff and Minkowski dimensions agree.

Example 1.3.4 Now we consider some sets in the plane. Suppose A is a b × b