This work is licensed under the Creative Commons Attribution 4.0 International License.
Besicovitch cascades
Let 𝕋 = ℝ/ℤ be a circle of length 1, Tρ : 𝕋 → 𝕋 be an irrational circle rotation
{T_\rho }x = x + \rho\,\,\,\, (\bmod 1),
and let f : 𝕋 → ℝ be a continuous function with zero mean. We consider a cylindrical transformation Tρ,f : 𝕋 × ℝ → 𝕋 × ℝ with a cocycle f :
{T_{\rho ,f}}(x,y) = ({T_\rho }x,y + f(x)),
which is a skew product over the circle rotation. (D.V. Anosov [1] also called it a cylindrical cascade.) The iterations of a point (x,y) are described by
T_{\rho ,f}^{r}(x,y)=(T_{\rho }^{r}x,y+{{f}^{r}}(x)),\quad r\in \mathbb{Z},
where fr (x) is a Birkhoff sum
f ^{r}(x)=\begin{cases}f(x)+f(T_{\rho}x)+\dots+f(T_{\rho}^{r-1}x) &\text{for } r>0,\\0\quad &\text{for } r=0,\\ -f(T_{\rho}^{-1}x)-\dots - f (T_{\rho}^{r+1}x)- f (T_{\rho}^{r}x) &\text{for } r<0.\end{cases}
Cylindrical transformation was studed by H. Poincaré [14] as a model for flat transformations. In particular, he studied the ω-limit points of orbits. Using these mappings, he, as well as other authors, constructed examples of flows in spaces of higher dimension with various topological properties. Cylindrical transformation have various applications in ergodic theory: see, for example, [13], [2], [15].
It turned out that there are cascades having unbounded orbits, and there are those only with bounded ones. If f is a coboundary over Tρ in the class of continuous functions, i.e., there exists F ∈ C(𝕋) such that
f(x) = F({T_\rho }x) - F(x)
for all x, then the orbit of each point (x0,y0) is contained within a closed invariant curve
\left\{ (x,y):y={{y}_{0}}+F(x)-F({{x}_{0}}),\quad x\in \mathbb{T} \right\},
and therefore it is bounded. In this case, the cylinder 𝕋 × ℝ splits into such curves, and Tρ,f is isomorphic to the rotation of the cylinder around the axis.
Later L.G. Shnirel’man [16] and A.S. Besicovitch [3] found examples of topologically transitive cylindrical cascades, i.e., cascades having dense orbits in the cylinder (such orbits are also called topologically transitive). At that time, they did not know whether all the orbits in this case could be topologically transitive.
In 1955 W.H. Gottschalk, G.A. Hedlund [8] showed that Tρ,fis topologically transitive if and only if f is not a coboundary over Tρand has zero mean.
The coboundaries and the cohomological (or homological in earlier terminology) equation (1) are of particular importance in the theory of dynamical systems. They are used to establish isomorphism of skew products, special flows, to study the spectral properties of metric automorphisms, etc. Cohomological equations can be considered in various classes of functions (see, for example, [1]). The Birkhoff sums for coboundaries have a convenient description
{f^r}(x) = F(T_\rho ^rx) - F(x).
In 1951 A.S. Besicovitch [4] proved that a cylindrical cascade cannot be minimal, it has nontransitive orbits, and for any irrational circle rotation Tρ, there exists a continuous f such that Tρ,f is topologically transitive and has discrete orbits (which are closed invariant sets). The point (x,y) ∈ 𝕋 × ℝ has a discrete orbit if and only if lim
\mathop {\lim }\limits_{|r| \to + \infty } {f^r}(x) = \infty .
The condition
\int_{\mathbb{T}}{}f(x)dx=0
is essential. Otherwise, by ergodic theorem, fr(x) → ∞ as r → ± ∞, and so each orbit is discrete.
The presence of discrete orbits, as well as the property of a function to be a coboundary, substantially depends on the properties of ρ. Recall the basic notions. Every irrational ρ ∈ (0,1) may be represented as an infinite continued fraction
\rho = {1 \over {{k_1} + {1 \over {{k_2} + \cdots }}}},
or briefly,
\rho = [{k_1},{k_2}, \ldots ],
and the natural numbers kn are called partial quotients. The fraction pn/qn = [k1,..., kn] is called the convergent of continued fraction. It is known [9], that the convergents pn/qn are determined recursively:
\matrix{ {{p_{n + 1}} = {k_{n + 1}}{p_n} + {p_{n - 1}},\;n \ge 1,\;{p_0} = 0,\;{p_1} = 1,} \hfill \cr {{q_{n + 1}} = {k_{n + 1}}{q_n} + {q_{n - 1}},\;n \ge 1,\;{q_0} = 1,\;{q_1} = {k_1}.} \hfill \cr } ρ is called Diophantine if there exist C > 0, θ > 0 such that
{q_{n + 1}} < Cq_n^{1 + \theta }
for any n. Otherwise, it is called Liouville.
In 2010, K. Frączek and M. Lemańczyk [7] began to study the properties of a set of discrete orbits depending on the function f and the rotation number ρ. (Following them, a transitive cascade with discrete orbits is called the Besicovitch cascade, and the set B ⊂ 𝕋 × {0} of circle points having discrete orbits is called the Besicovitch set.)
The Besicovitch set B is invariant under Tρ, Tρ is uniquely ergodic with the only invariant Lebesgue measure, and, therefore, B has a null Lebesgue measure.
Obviously, if f has bounded variation, then Tf is not Besicovitch, because the sequence
T_{\rho ,f}^{{q_n}}(x,y)
is bounded for denominators qn of ρ, as ‖ fqn‖C(𝕋) ⩽ Var(f).
K. Frączek and M. Lemańczyk showed that the Besicovitch cascade with continuous function f can be constructed for any transitive Tρ. For ρ satisfying some Diophantine condition, the γ-Hölder function f was obtained, so that Tρ,f is Besicovitch. In this construction, γ depends on the Diophantine parameter and γ < 1/2 in any case. They also showed that, under additional conditions, the Hausdorff dimension of the Besicovitch set can be at least 1/2.
Thus, it was established in [7] that both the admissible degree of continuity of the function and the Hausdorff dimension of the Besicovitch set depend on the properties of rotation.
We also note the result of E. Dymek [5], who showed that for any irrational ρ, one can construct a continuous cascade for which the Besicovitch set has a Hausdorff dimension 1.
A number of examples were constructed in [11], [12], demonstrating a closer relationship between the Hölder exponent of the function f and the obtained estimate of the Hausdorff dimension for the Besicovitch set. In [11], using angles with property
{q_{n + 1}} \asymp q_n^{1 + \theta }
, θ > 0, it was shown that for any γ ∈ (0,1) and any ɛ > 0, there exist a γ-Hölder function f and a circle rotation Tρsuch that the cylindrical transformation Tρ,fis Besicovitch, and the Hausdorff dimension of the Besicovitch set in the circle is greater than 1 − γ− ɛ.
In [12], using angles with relatively slowly varying, but infinitely large partial quotients, it was managed to achieve inequality dimH(B) ⩾ 1 − γ and also to construct the Besicovitch cascade with a function that is γ-Hölder with any exponent γ ∈ (0,1).
Here we prove the following theorem.
Theorem 1
For any γ ∈ (0,1) and any ɛ > 0, there exists a γ-Hölder function f and a circle rotation Tρwith bounded quotients such that the cylindrical transformation Tρ,fis Besicovitch, and the Hausdorff dimension of the Besicovitch set in the circle is greater than 1 − γ− ɛ.
Before proceeding to the proof of the theorem, we formulate two problems.
Problem 1
Is 1 − γ the upper bound for dimH(B)? It is unknown whether there exists a γ-Hölder Besicovitch cascade, for which the Hausdorff dimension of the Besicovitch set is greater than 1 − γ.
Problem 2
[7] Is it possible to construct a Besicovitch cascade with Hölder function over Liouville rotation of a circle?
The next three sections are devoted to the proof of the theorem.
The main construction
The construction is based on the design proposed in [11], but the operation with rotations having bounded partial quotients required a slight modification and more subtle estimates.
So, let ρ = [k1,...,kn,...] be an irrational number, and {pn/qn} be the sequence of convergents to ρ. Let δn = |ρ − pn/qn|. It is well known that 1
{1 \over {{q_n}({q_{n + 1}} + {q_n})}} < {\delta _n} < {1 \over {{q_n}{q_{n + 1}}}}
and
\rho = {{{p_n}} \over {{q_n}}} + {( - 1)^n}{\delta _n}.
We will construct the γ-Hölder cocycle f as the sum of a series of Lipschitz functions fn corresponding to fractions pn/qn:
f = \sum\limits_{n = 1}^\infty {f_n},
where fn is 1/qn–periodical function defined by expression
f_n(x)=\begin{cases}\frac{a_n}{2\delta_n}(x+\delta_n)^{2}, & x\in [-\delta_n,0],\\\frac{a_n\delta_n}{2}+a_nx, & x\in \left[0, \frac{1}{2t_nq_n}-\delta_n\right],\\\frac{a_n}{2t_nq_n}-\frac{a_n\delta_n}{4}-\frac{a_n}{\delta_n}\left( x-\frac{1}{2t_nq_n}+\frac{\delta_n}{2}\right)^{2},& x\in \left[ \frac{1}{2t_nq_n}-\delta_n, \frac{1}{2t_nq_n}\right],\\f_n\left( \frac{1}{t_nq_n}-\delta_n-x\right),& x\in \left[ \frac{1}{2t_nq_n}, \frac{1}{t_nq_n}\right],\\0, & x\in \left[ \frac{1}{t_nq_n}, \frac{1}{2q_n}-\delta_n\right],\\-f_n\left(x-\dfrac{1}{2q_n}\right), & x\in \left[ \frac{1}{2q_n}-\delta_n, \frac{1}{q_n}-\delta_n\right].\end{cases}
In this expression, in addition to the rotation number parameters, the quantities tn and an are used. We assume, that for any number n{t_n} > 2.
1/tn shows, what part of the period is occupied by the support of fn, and an > 0 is the Lipschitz constant for fn.
Note. Since the addition of the coboundary to the function
f = \sum\limits_{i = 1}^\infty {f_i}
does not change the Besikovich set, we can consider the terms and the convergents pn/qn, and also estimate the partial sums
\sum\limits_{i = {n_0}}^n f_i^r
, starting from some number n0. For simplicity, in this case we shift the numbering, assuming n0 = 1.
If an ⩽ const(tnqn)1−γfor any n and some γ ∈ (0,1), then the series(4)converges, and the function f is γ–Hölder.
The detailed proof of this lemma is given in [12]. The more general construction of functions with various continuity properties was also formulated and justified there. Earlier such construction of «almost Lipschitz function» was used by the author in [10].
The convergence of series (4) follows from the inequality
\parallel {f_n}{\parallel _C} \leqslant {{\rm const} \over {{{({t_n}{q_n})}^\gamma }}}
and the fact that denominators qn grow not slower than exponentially.
In this article we set
{a_n} = ({t_n}{q_n}{)^{1 - \gamma }}.
The conditions of the next lemma are modified compared with those in the previous papers [11] and [12]. These changes are adapted to dealing with bounded partial quotients.
Lemma 3
If the sequences {qn}, {tn} and {an} satisfy the conditions{{{q_{n + 1}}} \over {{t_n}{q_n}}} = m\geqslant 6,\quad m \in \mathbb{N},{{{a_n}} \over {{t_n}}} \to + \infty ,{{{a_n}{q_{n + 1}}} \over {t_n^2{q_n}}}:{{{a_{n + 1}}} \over {{t_{n + 1}}}} < {1 \over 6},then for any x ∈ D+\mathop {\lim }\limits_{r \to + \infty } {f^r}(x) = + \infty ,\quad \mathop {\lim }\limits_{r \to - \infty } {f^r}(x) = - \infty .For x ∈ D−, the situation is inverse.
Proof
1. At first, we will try to describe the mechanism of «pushing» the orbit of a point to infinity. The terms making up f, in turn, «are responsible» for the growth of Birkhoff sums on the set D+. Due to the good agreement with the circle rotation, the Birkhoff sum
f_n^r(x)
of the n-th term for
x \in G_n^ +
to a moment about r = qn/4 grows up to a level of the order an/tn, and at a moment of the order of
r = {{{q_{n + 1}}} \over {{t_n}}}
ceases to grow at a level no more than
{L_n} \approx {{{a_n}{q_{n + 1}}} \over {4t_n^2{q_n}}}
, and this value keeps constant up to the moment of at list qn+1/4, when the next term increases enough to «overtake» the current one. This fact explains the second and third assumptions of the lemma.
After the moment about r = qn+1/2, the Birkhoff sum
f_n^r(x)
decreases and may become negative, but the growth of the sums fr on the set
G_{n + 1}^ + \cap G_n^ +
is provided by the next summand
f_{n + 1}^r
. As each function fn is a coboundary, we can control the values of all previous terms, and the following terms are positive on D+.
2. We will show that fn is a coboundary. For this let us define a 1/qn–periodic function Fn : 𝕋 → ℝ such that
{f_n}(x) = {F_n}({T_\rho }x) - {F_n}(x).
We put
\begin{align*}\widehat{F}_n(x)=\begin{cases}\dfrac{a_n}{2\delta_n}x^{2}, & x \in \left[0, \dfrac{1}{2t_nq_n}\right],\\L_n-\widehat{F}_n\left(\dfrac{1}{t_nq_n}-x\right), & x\in \left[ \dfrac{1}{2t_nq_n}, \dfrac{1}{t_nq_n}\right],\\L_n, & x\in \left[ \dfrac{1}{t_nq_n}, \dfrac{1}{2q_n}\right],\\L_n-\widehat{F}_n\left(x-\dfrac{1}{2q_n}\right) ,& x\in \left[ \dfrac{1}{2q_n}, \dfrac{1}{q_n}\right],\end{cases}\end{align*}
where
{L_n} = 2{\widehat F_n}(1/(2{t_n}{q_n})) = {{{a_n}} \over {4{\delta _n}t_n^2q_n^2}}.
It easy to verify that
{f_n}(x) = {\widehat F_n}(x + {\delta _n}) - {\widehat F_n}(x).
Then we put
\begin{align*}{F}_n(x)=\begin{cases}\widehat{F}_n(x) & \textrm{ for even }n,\\L_n-\widehat{F}_n(x+\delta_n) & \textrm{ for odd }n.\end{cases}\end{align*}
For even n, when ρ = pn/qn + δn, we immediately have (11). For odd n, when ρ = pn/qn − δn,
{F_n}(x + \rho ) - {F_n}(x) = - {\widehat F_n}(x + {\delta _n} + \rho ) + {\widehat F_n}(x + {\delta _n}) = {\widehat F_n}(x + {\delta _n}) - {\widehat F_n}(x + {p_n}/{q_n}) = {f_n}(x).
For the constant Ln, by (3) and (8), the estimates
{m \over 4} \cdot {{{a_n}} \over {{t_n}}} < {L_n} < {m \over 4} \cdot {{{a_n}} \over {{t_n}}}\left( {1 + {{{q_n}} \over {{q_{n + 1}}}}} \right).
are valid.
3. Since for any integer r and any x ∈ 𝕋
f_n^r(x) = {F_n}(x + r\rho ) - {F_n}(x),\quad 0 \leqslant {F_n}(x) \leqslant {L_n},
we have
{\left\| {f_n^r} \right\|_C} \le {L_n}.
Moreover,
\left| {f_n^r(x)} \right| \leqslant {3 \over 4}{L_n}\quad {\rm{for}}\;{\rm{any}}\;x \in D_1^ \pm \;{\rm{and}}\;{\rm{any}}\;r \in \mathbb{Z},n \in \mathbb{N}.
Indeed, if
x \in D_1^ \pm
, then
x \in G_n^ \pm
, hence
{1 \over 4}{L_n} \leqslant {F_n}(x) \leqslant {3 \over 4}{L_n}
, and so
0 - {3 \over 4}{L_n} \leqslant f_n^r = {F_n}({T_\rho }x) - {F_n}(x) \leqslant {L_n} - {1 \over 4}{L_n}.
It follows from (8) and (10) that
{{{a_n}} \over {{t_n}}}:{{{a_{n + 1}}} \over {{t_{n + 1}}}} < {1 \over {6m}},
therefore, by (13), we obtain
{L_1} + \ldots + {L_n} < {m \over 4}\sum\limits_{i = 1}^n {{{a_i}} \over {{t_i}}}\left( {1 + {{{q_i}} \over {{q_{i + 1}}}}} \right) < {m \over 4}\left( {1 + {1 \over {{{\overline K }_n}}}} \right){{6m} \over {6m - 1}}{{{a_n}} \over {{t_n}}},
where, according to (5) and (8),
{\overline K _n} = \mathop {\min }\limits_{i \leqslant n} {{{q_{i + 1}}} \over {{q_i}}} > 12.
Given the last three inequalities, we obtain
{L_1} + \ldots + {L_n} < {m \over 3} \cdot {{{a_n}} \over {{t_n}}}.
4. Now let be
x \in G_n^ +
. Then
\quad f_n^{r + 1}(x)\geqslant f_n^r(x)\geqslant 0,\quad f_n^{ - r}(x) \leqslant 0\quad {\rm for}\, {\rm all}\, r \in \left[ {0,{q_{n + 1}}/4} \right],\quad |f_n^{ \pm r}(x)| \ge {{{a_n}} \over {18{t_n}}}\quad {\rm for}\, {\rm all}\, r \in \left[ {{q_n}/4,{q_{n + 1}}/4} \right].
From (16) it follows that if
x \in D_1^ + = \bigcap\nolimits_{n = 1}^\infty G_n^ +
, then for all r ∈ [0,qn+1/4]
f_m^r(x)\geqslant 0\quad {\rm{for}}\quad m\geqslant n.
To verify (16), note that by the periodicity of function fn and the set
G_n^ +
,
{f_n}(x + j/{q_n} + \rho ) = {f_n}(x + {( - 1)^n}{\delta _n}),
therefore, we can assume that x ∈ [1/(4tnqn),3/(4tnqn)]. For even n, and given that (qn+1/4)δn < 1/(4qn), tn > 2 and
r \leqslant {{{q_{n + 1}}} \over 4} - 1
,
{1 \over {4{t_n}{q_n}}} \leqslant x + r{\delta _n} \leqslant {3 \over {4{t_n}{q_n}}} + r{\delta _n} < {1 \over {2{q_n}}} - {\delta _n},
so all steps
T_\rho ^rx
within
r \leqslant {{{q_{n + 1}}} \over 4} - 1
occur in the region where fn is non-negative, which implies (16) and non-decreasing Birkhoff sums
f_n^r(x)
with increasing r in the indicated range.
The cases of odd n and r < 0 are similar.
To get (17), we consider r = ⎡qn/4 ⎤. As above, we can assume that x ∈ [1/(4tnqn),3/(4tnqn)], and
f_n^r(x) = \sum\limits_{k = 0}^{r - 1} {f_n}(x + {( - 1)^n}k{\delta _n}).
By (3) and (8), points of the form {x + (−1)nkδn, 0 ⩽ k ⩽ qn/4} fill evenly a segment of length shorter than 1/(4mtnqn) in the segment
\left[ {{{1 - 1/m} \over {4{t_n}{q_n}}},{{3 + 1/m} \over {4{t_n}{q_n}}}} \right]
, and therefore
f_n^r(x)\geqslant \left( {1 - {1 \over {2m}}} \right) \cdot {{{a_n}} \over {16{t_n}}}.
Roughing up the estimate, we obtain (17).
5. Now we can estimate Birkhoff sum fr(x) for
x \in D_1^ + \subset G_n^ +
and r ∈ [qn/4, qn+1/4]:
\matrix{ {{f^r}(x) = \sum\limits_{k = 1}^{n - 1} f_k^r(x) + f_n^r(x) + \sum\limits_{k = n + 1}^\infty f_k^r(x) > \quad \left\{ {{\rm by}\,(13),(18)} \right\}} \hfill \cr \,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,{ > - {3 \over 4}\left( {{L_1} + \ldots + {L_{n - 1}}} \right) + f_n^r(x) > \quad \left\{ {{\rm by}\,(15),(17)} \right\}\quad > - {m \over 4} \cdot {{{a_{n - 1}}} \over {{t_{n - 1}}}} + {1 \over {18}}{{{a_n}} \over {{t_n}}}.} \hfill \cr }
According to (8) and (10), we have
m \cdot {{{a_{n - 1}}} \over {{t_{n - 1}}}} < {{{a_n}} \over {6{t_n}}}
, so
{f^r}(x) > {{{a_n}} \over {{t_n}}}\left( {{1 \over {18}} - {1 \over {24}}} \right) = {1 \over {72}} \cdot {{{a_n}} \over {{t_n}}} \to \infty
for n → ∞, and therefore, for r → +∞.
The cases r → − ∞ and
x \in D_1^ -
are similar. Cases
x \in D_s^ \pm
, s > 1 require a shift in the numbering of terms.
Lemma 3 is proved.
Rotation of the circle
We consider an irrational number ρ with constant, starting from some position, partial quotients
\rho = [Q,K,K, \ldots ],
where Q is a rational number. In the standard representation with integer partial quotients, this means that partial quotients kn = K for all numbers n large enough. Thus, we define the sequence {qn} (and also {tn} and {an}). After that we fix the Hölder constant γ ∈ (0,1) and the constant
m\geqslant 6
appearing in the construction of f. It will give us, by (8),
{t_n} = {{{q_{n + 1}}} \over {m{q_n}}}
. To completely determine the cylindrical cascade Tρ,f, we define
{a_n} = {\left( {{t_n}{q_n}} \right)^{1 - \gamma }}.
So, we have three parameters K, γ and m, and we substitute them to the conditions of Lemma 3.
We have, starting from some number n,
{p_{n + 1}} = K{p_n} + {p_{n - 1}},\quad {q_{n + 1}} = K{q_n} + {q_{n - 1}}.
Since the sequence {qn} satisfies the difference equation, for n large enough we have
{q_n} = {C_1}{\lambda ^n} + {C_2}{( - \lambda )^{ - n}},
where
\lambda = {K \over 2} + \sqrt {{{{K^2}} \over 4} + 1} \approx K + {1 \over K}
is the root of characteristic equation λ2 − λ − 1 = 0. Also we have
[K,K,K, \ldots ] = {1 \over \lambda },\quad \rho = [Q,K,K, \ldots ] = {1 \over {Q + 1/\lambda }}.
We say that an ≈ bn if an = bn (1 + O(1/λ−2n)) for n → ∞. In this notation
{q_n} \approx C{\lambda ^n},\quad \quad {{{q_{n + 1}}} \over {{q_n}}} \approx \lambda .
If the parameters γ, m, K are such thatm\geqslant 6,\quad {\lambda ^{1 - \gamma }} > 6m,then the conditions of Lemma 3 are satisfied.
Proof
We used the condition (8) to determine tn. According to (24), the condition (10) of Lemma 3 holds for n large enough. By (21), the inequality tn > 6 holds for n large enough. Also, by (23), the condition (9), i. e.,
{{{a_n}} \over {{t_n}}} \to \infty
for n → ∞, is satisfied.
Hausdorff dimension of Besicovitch Set
In this section, we estimate the Hausdorff dimension of the Besicovitch set B, i. e., set of points on the circle 𝕋 × ℝ having discrete orbits. In section 2, the subset D ⊂ B of Besicovitch set was constructed (see (6)):
\mathop {\lim }\limits_{r \to \pm \infty } {f^r}(x) = \infty \quad {\rm{for\,any}}\,x \in D.
Thus, the lower bound for dimHD is also for dimHB. As for the upper bound for dimHD, the paper does not prove that it is such for dimHB, but such a hypothesis exists.
D is the union of a countable set of sets of Cantor type. It is sufficient to estimate
\mathop {\dim }\nolimits_H (D_1^ + )
because
\mathop {\dim }\nolimits_H (D) = \mathop {\dim }\nolimits_H (D_1^ + ).
Recall, that
D_1^ + = \bigcap\limits_{n = 1}^\infty G_n^ +
(see (6)). Let us call the segments of
G_n^ +
as n-th level segments.
We get the upper bound for dimHD by definition (see [6], section 2.1). Define the covers Wn of
D_1^ +
by segments inductively.
{W_1} = G_1^ +
(we consider the cover as the union of segments). The cover Wn consists of those segments of
G_n^ +
that intersect Wn−1 in more than one point. The length of each of them is equal to 1/(2tnqn). By virtue of (8), each nth-level segment may intersect m/2 or m/2 + 1 segments of (n + 1)-th level. If mn is the number of nth–level segments included into Wn, than
{m_n} \leqslant \left( {{m \over 2} + 1} \right){m_{n - 1}},
, or
{m_n} \leqslant {m_1}{\left( {{m \over 2} + 1} \right)^{n - 1}}.
Thus, we have the cover of
D_1^ +
by segments of length
1/(2{t_n}{q_n}) \approx {C \over m} \cdot {\lambda ^{n + 1}}
(by (20) and (5)). Then we have the estimate
H_n^d = {m_n} \cdot 1/(2{t_n}{q_n}{)^d} < {\rm const}{\left( {{{{m \over 2} + 1} \over {{\lambda ^d}}}} \right)^{n - 1}}.
If
{\lambda ^d} > \left( {{m \over 2} + 1} \right)
, then
\mathop {\lim }\limits_{n \to \infty } H_n^d = 0
. Therefore, for Hausdorff dimension
d = \mathop {\dim }\nolimits_H D_n^ +
, the inequality
{\lambda ^d} \leqslant \left( {{m \over 2} + 1} \right)
is necessary, and we get the upper bound for Hausdorff dimension.
We obtain the lower bound for Hausdorff dimension by the method proposed in [6], Example 4.6. Now we denote by sn the minimal number of nth-level segments, contained in one (n − 1)th-level segment. Let ɛn be the minimal distance between n-th level segments. Then
\mathop {\dim }\nolimits_H (D_1^ + )\geqslant \mathop {\liminf}\limits_{n \to \infty } {{\ln ({s_2} \ldots {s_n})} \over { - \ln ({s_{n + 1}}{\varepsilon _{n + 1}})}}.
By (8), each n-th level segment contains entirely m/2 or (m/2 − 1) segments of the form [(i − 1) /qn+1, i/qn+1], therefore, each nth-level segment entirely contains at least (m/2 − 1) (n + 1)-th level segments, where from sn ⩾ m/2 − 1, we have
{\varepsilon _n} = {1 \over {{q_n}}}\left( {1 - {1 \over {2{t_n}}}} \right)
. Substituting these inequalities and (20) to (29), we obtain
\mathop {\dim }\nolimits_H (D_1^ + )\geqslant \mathop {\lim }\limits_{n \to \infty } {{(n - 1)\ln \left( {{m \over 2} - 1} \right)} \over {\ln {q_{n + 1}} + \ln {s_{n + 1}} - \ln \left( {1 - {1 \over {2{t_n}}}} \right)}} = {{\ln \left( {{m \over 2} - 1} \right)} \over {\ln \lambda }},
and we get the lower bound in (26).
Lemma 6
For given γ ∈ (0,1) and any ɛ > 0, there exist an even integer m ⩾ 6, and the constant partial quotients K defining ρ, such that{\lambda ^{1 - \gamma }} > 6m\quad {\rm{and}}\quad {{\ln \left( {{m \over 2} - 1} \right)} \over {\ln \lambda }} > 1 - \gamma - \varepsilon .
Proof
Recall, that
\lambda = {K \over 2} + \sqrt {{{{K^2}} \over 4} + {1 \over 4}} \approx K + {1 \over K}
. Since
\mathop {\lim }\limits_{\lambda \to + \infty } {{{\lambda ^{1 - \gamma }}/6} \over {2\left( {{\lambda ^{1 - \gamma - \varepsilon }} + 1} \right)}} = + \infty ,
the double inequality
{\lambda ^{1 - \gamma }}/6 < m < 2\left( {{\lambda ^{1 - \gamma - \varepsilon }} + 1} \right),
which is equivalent to (30), is resolvable in the class of even m for any λ (and hence K) large enough.
This lemma completes the proof of Theorem 1.
In fact, for given γ ∈ (0,1) and ɛ > 0, we choose parameters m ⩾ 6 and K satisfying (30). Using these parameters and some rational Q, we define ρ = [Q,K,K,...] and thus a sequence of convergents {pn/qn}. Now, according to (8), the sequences {tn} and {an = (tnqn)1−γ} are also defined.
Then, using the main construction, we define the function f and the set D.
According to Lemma 4, all the conditions of Lemma 3 are satisfied, therefore all points x ∈ D run away to infinity under iterations of the cylindrical cascade Tρ,f. So Tρ,f is a Besicovitch cascade, and D is a Besicovitch subset. By Lemma 2, f is γ-Hölder.
As m and K satisfy (25), from Lemma 5 we get the estimate
\mathop {\dim }\nolimits_H B > 1 - \gamma - \varepsilon .