www.imstat.org/aihp 2014, Vol. 50, No. 4, 1371–1384
DOI:10.1214/13-AIHP556
© Association des Publications de l’Institut Henri Poincaré, 2014
Hausdorff dimension of affine random covering sets in torus
Esa Järvenpää
a, Maarit Järvenpää
a, Henna Koivusalo
a, Bing Li
b,aand Ville Suomala
aaDepartment of Mathematical Sciences, P.O. Box 3000, 90014 University of Oulu, Finland.
E-mail:esa.jarvenpaa@oulu.fi;maarit.jarvenpaa@oulu.fi;henna.koivusalo@oulu.fi;ville.suomala@oulu.fi
bDepartment of Mathematics, South China University of Technology, Guangzhou, 510641, P.R. China. E-mail:libing0826@gmail.com Received 2 July 2012; revised 20 February 2013; accepted 27 February 2013
Abstract. We calculate the almost sure Hausdorff dimension of the random covering set lim supn→∞(gn+ξn)ind-dimensional torusTd, where the setsgn⊂Tdare parallelepipeds, or more generally, linear images of a set with nonempty interior, andξn∈Td are independent and uniformly distributed random points. The dimension formula, derived from the singular values of the linear mappings, holds provided that the sequences of the singular values are decreasing.
Résumé. Nous calculons presque sûrement la dimension de Hausdorff de l’ensemble de recouvrement aléatoire lim supn→∞(gn+ ξn)dans le toreTdde dimensiond, oùgn⊂Td sont des parallélépipèdes, ou plus généralement, des images linéaires d’un en- semble d’intérieur non vide etξn∈Tdsont des points aléatoires indépendants et uniformément distribués. La formule de dimen- sion, exprimée en fonction des valeurs singulières des applications linéaires, est valable à condition que la suite de ces valeurs singulières soit décroissante.
MSC:60D05; 28A80
Keywords:Random covering set; Hausdorff dimension; Affine Cantor set
1. Introduction
Given a sequence of positive numbers(ln)and a sequence of independent random variables(ξn)uniformly distributed on the circleT1=R/Z, definethe random covering setEas follows:
E=
x∈T1|x∈ [ξn, ξn+ln]for infinitely manyn
=lim sup
n→∞ [ξn, ξn+ln].
Denoting the Lebesgue measure by L and using the Borel–Cantelli lemma and Fubini’s theorem, it follows that, almost surely, the following dichotomy holds:
L(E)=
0, when∞
n=1ln<∞, 1, when∞
n=1ln= ∞, (1.1)
that is, almost all or almost no points of the circle are covered, depending on whether or not the series of the lengths of the covering intervals diverges.
The case of full Lebesgue measure has been extensively studied. It was a long-standing problem to find conditions on(ln)guaranteeing that the whole circle is covered almost surely, that is,
P
E=T1
=1. (1.2)
This problem, known in literature as the Dvoretzky covering problem, was first posed by Dvoretzky [5] in 1956. After substantial contribution of many, including Kahane [18], Erd˝os [7], Billard [3] and Mandelbrot [24], the full answer was given by Shepp [29] in 1972. He proved that (1.2) holds if and only if
∞ n=1
1
n2exp(l1+ · · · +ln)= ∞,
where the lengths(ln)are in decreasing order. After this, a natural problem, raised by Carleson (private communication to Kahane), is to describe the growth of the covering number of a given pointx∈T1, that is, to study the asymptotic behaviour of the sums
CN(x)= N
n=1
χ[ξn,ξn+ln](x),
whereχAis the characteristic function of a setA. Obviously, the expectationE(CN(x))=N
n=1ln. In the caseln=γn withγ >1, Fan and Kahane [10] proved that almost surely the order of the covering numberCN(x)is logNfor every x∈T1, meaning that for sufficiently largeN
AγlogN≤min
x∈T1CN(x)≤max
x∈T1CN(x)≤BγlogN
with positive and finite constantsAγ andBγ. Furthermore, Fan [9] verified that the set Fβ=
x∈T1 lim
N→∞
CN(x) N
n=1ln
=β
has positive Hausdorff dimension for a certain interval ofβ >0 in the caseln=γn withγ >0. For generalln, Barral and Fan [2] answered Carleson’s problem by identifying three kinds of phenomena depending whether the index
¯
γ =lim supN→∞
N n=1ln
−loglN is zero, positive and finite or infinite. More precisely, when γ¯ =0, dimHFβ =1 almost surely for allβ≥0, whenγ¯= ∞,F1=T1almost surely, and when 0<γ <¯ ∞, dimHFβ depends onβ. Here the Hausdorff dimension is denoted by dimH.
For the case of zero Lebesgue measure, the Hausdorff dimension ofEwas first calculated by Fan and Wu [12]
in the caseln=1/nα. When studying the Hausdorff measure and the large intersection properties ofEfor general ln, Durand [4] gave another, independent proof of the dimension result. According to [12] and [4], the almost sure Hausdorff dimension ofEis given by
dimHE=inf
t≥0 ∞
n=1
lnt <∞
=lim sup
n→∞
logn
−logln
, (1.3)
where the lengthsln are in decreasing order. In [4], the author also proved that the packing dimension ofEequals 1 almost surely. When considering the hitting probability property of the random setE, Li, Shieh and Xiao [22]
provided an alternative way to obtain the Hausdorff and packing dimension results under some additional conditions.
The result (1.3) can be also proven as a consequence of the mass transference principle due to Beresnevich and Velani [1] (see Proposition4.7). The fact that both packing and box counting dimensions are equal to 1 almost surely follows sinceEis almost surely a denseGδ-set inT1(see [19], Chapter 5, Proposition 11, and [27], Section 2).
In this paper we study random covering sets ind-dimensional torusTd. Letting(gn)be a sequence of subsets of Td and letting(ξn)be a sequence of independent random variables, uniformly distributed onTd, define the random covering set by
E=lim sup
n→∞ (gn+ξn)= ∞ n=1
∞ k=n
(gk+ξk).
Notice that a counterpart of (1.1) is easily obtained, that is, almost surely L(E)=
0, when∞
n=1L(gn) <∞, 1, when∞
n=1L(gn)= ∞, whereLis the Lebesgue measure onTd.
On thed-dimensional torus the Dvoretzky covering problem has been studied by El Hélou [6] and Kahane [20]
among others. In [20] Kahane gave a complete solution for the problem when the setsgn are similar simplexes (see also Janson [16]). However, in the general case the covering problem has not been completely solved.
For an overview on the research on random covering sets and related topics, we refer to [19], Chapter 11, the survey [21] and the references therein. Here we only mention a few variations on the classical random covering model. For example, Hawkes [13] considered under which conditions all the points inK⊂T1are covered with probability one (or zero). Mandelbrot [25], in turn, introduced Poisson covering of the real line (see also Shepp [28]). In general metric spaces, the random coverings by balls have been studied by Hoffman-Jörgensen [15]. Recent contributions to the topic include various types of dynamical models, see Fan, Schmeling and Troubetzkoy [11], Jonasson and Steif [17] and Liao and Seuret [23].
We address the question of determining the analogue of (1.3) in higher dimensional case. In [12] the method is strongly adapted to the 1-dimensional case whereas the argument based on the mass transference principle [1] can be carried through in any dimension provided that the setsgnare uniformly ball like (see Proposition4.7). Our main interest is the case where the setsgnare not uniformly ball like, and therefore, the mass transference principle cannot be applied. It turns out that almost surely the Hausdorff dimension of the covering set E is given in terms of the singular value functions of the linear mappings related to the system, see Theorem2.1.
To this end, in Section2we introduce our setting, state our main result and prove preliminary lemmas including the upper bound for the dimension. In Section3 we construct a random subset of the covering setE having large dimension with positive probability which, in turn, gives the lower bound of the dimension in Section4.
2. Preliminaries and statement of main theorem
Denote the closed ball of radiusrand centrexinRdbyB(x, r). LettingL:Rd→Rdbe a contractive linear injection, the imageL(B(0,1))of the unit ballB(0,1)is an ellipse whose semiaxes are non degenerated. The singular values 0< αd(L)≤ · · · ≤α1(L) <1 ofLare the lengths of the semiaxes ofL(B(0,1))in decreasing order. Given 0< s≤d, define thesingular value functionby
Φs(L)=α1(L)· · ·αm−1(L)αm(L)s−(m−1), wheremis the integer such thatm−1< s≤m.
We use the notationsTdfor thed-dimensional torus andLfor the Lebesgue measure onTd. Consider a probability space(Ω,A, P )and let(ξn)be a sequence of independent random variables which are uniformly distributed onTd, that is,(ξn)∗P =L, where(ξn)∗P is the image measure ofP underξn. Letting(gn)be a sequence of subsets ofTd, we use the notationGn for the random translatesGn=gn+ξn⊂Td and define the random covering set generated by(gn)by
E=Eω=lim sup
n→∞ Gn.
In this paper we consider the casegn=Π (Ln(R)), whereR⊂ [0,1]d has non-empty interior,Ln:Rd→Rd is a contractive linear injection for alln∈NandΠ:Rd→Tdis the natural covering map. Moreover, we assume that for alli=1, . . . , dthe sequence of singular valuesαi(Ln)decreases to 0 asntends to infinity. Defining
s0=inf
0< s≤d ∞
n=1
Φs(Ln) <∞
, (2.1)
with the interpretations0=d if∞
n=1Φd(Ln)= ∞, we are ready to state our main theorem.
Theorem 2.1. ForP-almost allω∈Ωwe have
dimHEω=s0. (2.2)
Theorem2.1is an immediate consequence of the following proposition concerning the case where each generating setgn is a rectangular parallelepiped inTd meaning that there exist a parallelepipedg˜n⊂Rd such thatgn=Π (g˜n).
In what follows rectangular parallelepipeds will consistently be called rectangles.
LetE(gn)=Eω(gn)be the covering set generated by a sequence(gn)of rectangles. For all rectanglesgand for all 0< s≤d define
Φs(g)=α1(g)· · ·αm−1(g)αm(g)s−(m−1),
where 0< αd(g)≤ · · · ≤α1(g) <1 are the lengths the edges ofgin decreasing order andmis the integer such that m−1< s≤m.
Proposition 2.2. Assume that(gn)is a sequence of rectangles such that for alli=1, . . . , d the sequence of lengths αi(gn)decreases to0asntends to infinity.Then almost surely
dimHE(gn)=s0(gn), (2.3)
where
s0(gn)=inf
0< s≤d ∞
n=1
Φs(gn) <∞
with the interpretations0=dif∞
n=1Φd(gn)= ∞.
We proceed by verifying first that Theorem2.1follows from Proposition2.2.
Proof of Theorem2.1as a consequence of Proposition2.2. Letting(Ln),RandEbe as in Theorem2.1, there are sequences(gn)and(gn)of rectangles such thatgn⊂Π (Ln(R))⊂gn, and moreover,αi(gn)=cαi(Ln)andαi(gn)= cαi(Ln)for alli=1, . . . , d. Here the constantsc andcare independent ofnandi. SinceE(gn)⊂E⊂E(gn)we have
dimHE gn
≤dimHE≤dimHE(gn).
Applying Proposition2.2to the sequences(gn)and(gn)and noting thats0(gn)=s0(gn)=s0, gives (2.2).
It remains to prove Proposition2.2. As the first step we verify the following lemma according to which the Haus- dorff dimension ofE(gn)is always bounded above bys0(gn). The proof is standard following, for example, the ideas in [8].
Lemma 2.3. Assume that(gn)ands0(gn)are as in Proposition2.2.Then for allω∈Ω we havedimHEω(gn)≤ s0(gn).
Proof. We may assume thats0(gn) < d. Lets0(gn) < s < d and letmbe the integer withm−1< s≤m. For each n∈Nwe estimate the number of cubes of side lengthαm(gn)needed to coverGn. By expanding the lastd−m+1 edges ofGnto lengthαm(gn)and by dividing the expanded rectangle to cubes of side lengthαm(gn), we end up with an upper bound
α1(gn) αm(gn)
+1
. . .
αm−1(gn) αm(gn)
+1
≤2m−1α1(gn)· · ·αm−1(gn)αm(gn)−m+1, where the integer part of anyx≥0 is denoted byx.
Recalling that for allN∈N E(gn)⊂ ∞
n=N
Gn,
gives the following estimate for thes-dimensional Hausdorff measure Hs(E)≤lim inf
N→∞
∞ n=N
2m−1√
dαm(gn)s
α1(gn)· · ·αm−1(gn)αm(gn)−m+1
=lim inf
N→∞
∞ n=N
2m−1(√
d)sΦs(gn)=0.
This implies that dimHE(gn)≤s0(gn).
We continue by proving two auxiliary results.
Lemma 2.4. Assume that(Ln)is a sequence of contractive linear injectionsLn:Rd→Rd.Lets0be as in(2.1)and letm−1< s0≤m.Defining for allm−1< s < s0
f (s):=lim sup
n→∞
logn
−logΦs(Ln), we havef (s) >1.
Proof. We will show thatf (s)≥1 for allm−1< s < s0andf is strictly decreasing. This clearly implies the claim.
Letm−1< s < s0. The fact that∞
n=1Φs(Ln)= ∞implies that for allε >0 there exists a subsequence(nk) such thatΦs(Lnk) > 1
n1k+ε for allk. From this we deduce thatf (s)≥1+1ε, and lettingεgo to 0 yieldsf (s)≥1.
Considerδ >0 such thatm−1< s+δ < s0. SinceΦs(Ln)≥αm(Ln)s we obtain Φs+δ(Ln)=Φs(Ln)αm(Ln)δ≤Φs(Ln)1+δ/s,
giving
f (s+δ)≤lim sup
n→∞
logn
(1+δ/s)(−logΦs(Ln))= f (s)
1+δ/s < f (s).
Hencef is strictly decreasing.
Remark 2.5. Lemma2.4holds for all0< s < s0,but this stronger claim is not necessary for our purposes.
Proposition 2.6. Assume that G⊂Td andL(G) >0.Letξ1, . . . , ξn be independent,uniformly distributed random variables onTd.Let
Mn=#
i∈ {1, . . . , n}|ξi∈G ,
where#Adenotes the number of elements in a setA.Then P
Mn≤1
2nL(G)
≤4(1−L(G)) nL(G) .
Proof. Denote byχAthe characteristic function of a setA. Calculating the first and second moments ofMngives E(Mn)=E
n
i=1
χ{ξi∈G}
=nL(G)
and E
Mn2
=E n
i=1
χ{ξi∈G}
2
=E n
i=1
χ{ξi∈G}+
j=i
χ{ξi∈G}χ{ξj∈G}
=nL(G)+ n2−n
L(G)2. From Chebyshev’s inequality we deduce
P
Mn≤1 2E(Mn)
≤P Mn−E(Mn) ≥1 2E(Mn)
≤4(E(Mn2)−E(Mn)2)
E(Mn)2 =4(1−L(G)) nL(G)
which completes the proof.
3. Construction of random Cantor sets
Let(gn)ands0(gn)be as in Proposition2.2. Consider an integermsuch thatm−1< s0(gn)≤m. For notational sim- plicity, we assume that 0 is a vertex of eachgn. Indeed, by choosing suitable deterministic translates, we find an iso- morphic probability space(Ω,A, P )where this is the case since the random variables(ξn)are uniformly distributed and the rectangles(gn)are deterministic. For eachn, letTn:Rd→Rdbe a linear map such thatΠ (Tn([0,1]d))=gn. Observe thatαi(Tn)=αi(gn)for alli=1, . . . , d. Letm−1< s < s0(gn). For the purpose of proving Proposition2.2 we construct in this section an eventΩ(∞)⊂Ω, having positive probability, and a random Cantor like setCω such thatCω⊂Eωfor allω∈Ω(∞). In Section4we prove that dimHCω≥salmost surely conditioned onΩ(∞).
Leta0= 12. Consider a sequence(al)of real numbers larger than 1/2 increasing to 1 with ∞
l=1 1
al <∞. By Lemma2.4, there exists a sequence(nk)of natural numbers satisfying
klim→∞
lognk
−logΦs(Tnk)=f (s) >1. (3.1)
Moreover, by considering a suitable subsequence of(nk), we may assume that for allk∈N diam(gnk)≤1
2(1−ak−1)αd(gnk−1), (3.2)
nkL(gnk−1)≥n(3k+f (s))/(2+2f (s)) and (3.3)
lognk≥nk−1, (3.4)
wheren0=0 andg0=Td. Notice that since the sequence(nk)is deterministic it is independent ofω∈Ω.
We proceed by constructing inductively a random nested sequence of finite collectionsCkof rectangles as follows:
LetC0= {Td}andN0=1. DefineN1= 12ad0n1andI (1,Td)= {1, . . . , N1}. For alli∈I (1,Td), letgi be a linear isometric copy ofgn1 contained ingi. The existence ofgi follows from the fact thatαj(gn1)≤αj(gi)for alli≤n1 andj=1, . . . , d. For eachi∈I (1,Td), setGi=gi+ξi. ThenGi⊂Gi. DefiningC1= {Gi|i∈I (1,Td)}, we have
G∈C1
G⊂
n1
i=1
Gi.
Furthermore, the collectionC1can be chosen for anyω∈Ω=:Ω(1)givingP (Ω(1))=q1withq1=1.
Assume that there exist events Ω(1), . . . , Ω(k−1) with P (k−1
j=1Ω(j )) =q1· · ·qk−1 such that for all ω∈ k−1
j=1Ω(j )there are collectionsC1, . . . ,Ck−1having the following properties for allj=1, . . . , k−1 1
2 3
ajd−1(nj−nj−1)L(gnj−1)≤Nj≤(nj−nj−1)L(gnj−1), whereNj=#Cj, (3.5)
G∈Cj
G⊂
G∈Cj−1
G, (3.6)
#
G ∈Cj|G ⊂G
= 1
2adj−1mjL(gnj−1)
for eachG∈Cj−1
wheremj=
nj−nj−1 Nj−1
, (3.7)
Cj is a finite collection of isometric copies ofgnj and (3.8)
G∈Cj
G⊂
nj
l=nj−1+1
Gl. (3.9)
We define an event Ω(k)such that P (k
j=1Ω(j ))=q1· · ·qk and for all ω∈k
j=1Ω(j )there is a collectionCk
satisfying (3.5)–(3.9). WriteCk−1= {G1, . . . ,GNk−1}and setmk= nkN−kn−k−11 . Forl=1, . . . , Nk−1, define random sets
I (k,Gl)= i∈
nk−1+1+(l−1)mk, . . . , nk−1+lmk
|ξi∈ak−1Gl ,
whereaGis the similar copy ofGwith similarity ratioaand with the same centre asG. Let Ω(k)=
ω∈Ω #I (k, G) > 1
2akd−1mkL(gnk−1)for allG∈Ck−1
and
qk=P
Ω(k)|Ω(1), . . . , Ω(k−1) .
Note thatqk>0. For eachG∈Ck−1we denote byI (k, G)the collection of the first12adk−1mkL(gnk−1)elements in I (k, G)and set
Ck=
Gi|G∈Ck−1, i∈I (k, G)
and Nk=#Ck,
whereGi=gi+ξi andgi is a linear isometric copy ofgnk contained ingi. (See Fig.1.) Observe thatNk is deter- ministic. As above,gi exists sinceαj(gnk)≤αj(gi)for allj=1, . . . , d andi≤nk. Clearly, (3.7) and (3.8) are valid forCk. Since, by inequality (3.2), we havegi+ξi⊂G∈Ck−1provided thatξi ∈ak−1G, property (3.6) holds forCk. Furthermore, the choices ofmkandI (k, Gl)imply (3.9). The choice ofmkgives
1 2
3
akd−1(nk−nk−1)L(gnk−1)≤Nk−1 1
2akd−1mkL(gnk−1)
=Nk
≤(nk−nk−1)L(gnk−1), and therefore, condition (3.5) is satisfied forCk. Finally,
P k
l=1
Ω(l)
=P
Ω(k)|Ω(1), . . . , Ω(k−1) P
k−1
l=1
Ω(l)
=q1· · ·qk.
Fig. 1. Construction ofCk.
LettingΩ(∞)=∞
n=1Ω(n), we haveP (Ω(∞))=∞
n=1qn. Define for allω∈Ω(∞) Cω=∞
n=1
G∈Cn
G⊂Eω.
Next we verify that the Cantor like setCω ⊂Eω exists with positive probability. We use the notationFk for the σ-algebra generated by the random variablesξ1, . . . , ξnk.
Proposition 3.1. With the above notation we haveP (Ω(∞)) >0.
Proof. We have qk=P
Ω(k)|Ω(1), . . . , Ω(k−1)
= 1
P (k−1 l=1Ω(l))P
Ω(k)∩
k−1
l=1
Ω(l)
= 1
P (k−1
l=1Ω(l))E
E(χΩ(k)χk−1
l=1Ω(l)|Fk−1)
= 1
P (k−1
l=1Ω(l))E χk−1
l=1Ω(l)E(χΩ(k)|Fk−1)
= 1 P (k−1
l=1Ω(l))E χk−1
l=1Ω(l)E(χ
G∈Ck−1{#I (k,G)>(1/2)ak−1d mkL(gnk−1)}|Fk−1)
≥ 1
P (k−1
l=1Ω(l))E
χk−1 l=1Ω(l)
1−
G∈Ck−1
E(χ{#I (k,G)≤(1/2)m
kL(ak−1gnk−1)} Fk−1)
,
and applying Proposition2.6hence gives qk≥1−Nk2−1 8(1−L(ak−1gnk−1))
(nk−nk−1)L(ak−1gnk−1)=:1−pk.
Inequalities (3.5) and (3.4), in turn, imply thatNk−1≤(nk−1−nk−2)L(gnk−2)≤nk−1≤lognk, and therefore, noting thatnk−nk−1≥12nkby (3.4) and using (3.3), we obtain
∞ k=1
pk≤ ∞ k=1
8(1−L(ak−1gnk−1))(lognk)2 (1/2)akd−1nkL(gnk−1) ≤
∞ k=1
8(lognk)2
(1/2)akd−1n(3k+f (s))/(2+2f (s))
<∞,
where the convergence follows since by (3.4) the sequence(nk)is growing exponentially fast. Lettingk0∈Nbe such thatpk<1 for allk≥k0, we have∞
k=1qk≥k0
k=1qk∞
k=k0+1(1−pk) >0.
Remark 3.2. The idea of finding a large-dimensional Cantor subset of the random covering set was already exploited in the dimension calculation of Fan and Wu[12]in the case ofT1.In their proof it is essential that the setsCω are homogeneous and the construction intervals are well-separated,which follows from well-known results on random spacings of uniform random samples[14].Structure of the set allows them then to directly estimate sizes of inter- sections of balls with the set Cω,giving the dimension bound from below.In our choice of the subsetCω,however, separation of the generating sets plays no role.Indeed,it is a well-known fact that for self-affine sets no separation condition guarantees the dimension formula.Also a direct estimate for measures of balls is probably hopeless.Instead a potential theoretic method based on a transversality argument is the key,see Lemma4.3below.In the implementation of this idea we need the assumption(3.8).
4. Dimension estimate
Using the notation introduced in Section3, we prove that fors < s0(gn)the event{ω∈Ω(∞)|dimHCω≥s}has pos- itive probability. To obtain the dimension bound, we use potential theoretic methods and define a measure supported onCωwith finites-energy. In what follows, we consider only the eventΩ(∞)and denote the expectation overΩ(∞) simply byE.
For any ω∈Ω(∞),k∈N andG∈Ck−1, let Mk=#I (k, G)= 12adk−1mkL(gnk−1)be the number of levelk construction rectangles contained in G. Notice that Mk is a deterministic number depending only on k. For later notational simplicity, we will relabel the random variablesξi using a deterministic tree structure.
For alll∈N, consider the setsJl= {i1. . . il|ij∈ {1, . . . , Mj}for allj∈ {1, . . . , l}}and defineJ=∞
l=0Jl, with the conventionJ0= {∅}. Fori,j∈J, denote byi∧jthe maximal common initial sequence ofiandjand letij∈Jbe the word obtained by juxtaposing the wordsiandj. Further, we denote by|i|the length ofi∈J, that is,|i| =lifi∈Jl. For eachl≤kandi∈Jl, define thecylinderof lengthland of depthkbyC(i, k)= {j∈Jk|i∧j=i}. Fori∈ {1, . . . , M1}, defineφi=ξi andG(i)=gi+φi and letTi be a linear map such that Π (Ti([0,1]d)=gi. Assume that we have defined the random variables φi and the rectangles G(i)∈Ck−1 for all i∈Jk−1. Let I (k, G(i))= {j1, . . . , jMk} whereji < ji+1 in the natural order given by the construction. For alli∈ {1, . . . , Mk}, defineφii =ξji,gii =gj
i
and G(ii)=gii +φii and let Tii be a linear map satisfying Π (Tii([0,1]d))=gii. Then det(Tn|i|)=L(G(i)) and Φs(Ti)=Φs(Tn|i|)for alli∈J. For notational purposes setG(∅)=Tdandφ∅=0. When necessary we viewTi as a map onTdby identifyingTdwith[0,1[d. Finally, fori1, . . . ,ik∈J, denote byF(i1, . . . ,ik)theσ-algebra generated by the events{ω∈Ω(∞)|G(il)=Ql for alll=1, . . . , k}, where eachQl⊂Td is an isometric copy ofgn|il|. Remark 4.1. Note that {φi|i ∈ C(j, k)} = {ξi|i ∈ I (k, G(j))} for any j ∈ Jk−1 and {φi|i ∈ Jk} = {ξi|i ∈
G∈Ck−1I (k, G)}.Let A⊂Td be a Borel set withL(A) >0.Since ξj is uniformly distributed onTd for givenj,
everyξjis uniformly distributed onAwhen conditioned on the eventξj∈A.Leti∈Nand letii∈Jk+1.By definition φii=ξj for somej∈ {nk+1, . . . , nk+1}withξj∈G(i),and hence the random variableφii is uniformly distributed onakG(i)when conditioned onφii=ξj and theσ-algebraF(i).Furthermore,
E
χ{φii∈A}|F(i)
=
nk+1
j=nk+1
E
χ{φii∈A}|F(i), φii=ξj E
χ{φii=ξj}|F(i)
=L(A∩akG(i)) L(akG(i))
nk+1
j=nk+1
E
χ{φii=ξj}|F(i)
=L(A∩akG(i)) L(akG(i)) .
Henceφii is uniformly distributed insideakG(i)when conditioned onF(i).Moreover,ifjsatisfiesj∧ii=ii,con- ditioning onF(i,j)instead ofF(i)does not change the uniform distribution of φii onakG(i),sinceξj and ξl are independent forj=l.Recall that even though the corner pointsφii andφihare independent fori=h,the rectangles G(ii)andG(ih)are not,since the orientation ofgii is determined by the indexji.
Lemma 4.2. The sequence of measuresμωl onTdgiven by
μωl =
i∈Jl(Ti +φi)∗L
Nl (4.1)
converges in weak∗-topology to a measureμω supported onCω.
Proof. By the Riesz representation theorem a weak∗-limitμω exists, if we prove that for all positive, continuous functionsf onTdthe sequence
fdμωl converges.
To that end, fix a positive, continuous functionf onTdandε >0. SinceTd is compact, there existsδ >0 with
|f (x)−f (y)|< εfor all|x−y|< δ. LetKbe so large that diam(gnK) < δ, and fixk≥K. Writeμωk as a sum of measuresμωi,kdefined by
μωk =
i∈JK
j∈C(i,k)
(Tj+φj)∗L
Nk =
i∈JK
μωi,k.
For alli∈JK, we haveμωi,k(G(i))=N1K =μωi,K(G(i))and sptμωi,k⊂G(i). Therefore,
fdμωk −
fdμωK ≤
i∈JK
G(i)
fdμωi,k−
G(i)
fdμωi,K ≤ε, since diamG(i)=diam(gnK) < δ. Thus sequence
fdμωl converges. The claim sptμω ⊂Cω holds since Cω is compact and sptμωl ⊂
G∈ClGfor alll.
Next we show that for alls < s0(gn)thes-energyIs(μω)= dμω|(x)dμx−y|sω(y) ofμω is finite almost surely. In the energy estimate we will make use of the following lemma [8], Lemma 2.2.
Lemma 4.3 (Falconer). Letsbe non-integral with0< s < d and letT:Rd→Rdbe an affine injection.Then there exists a number0< D0<∞,depending only ondands,such that
[0,1]d
dL(x)
|T (x)|s ≤ D0
Φs(T ).
Lemma 4.4. For alli,j∈Jandx, y∈Td we have
E
χG(j)(y)χG(i)(x)
≤ ∞
l=1
1 al
2d
det(Tn|i|)det(Tn|j|) det(Tn|i∧j|)2 E
χG(i∧j)(y)χG(i∧j)(x) .