The word
The main tool used in these areas to deal with fractals is the fractal dimension since it is their main invariant which throws some useful information about the complexity and irregularitires that a certain set presents once it has been explored with enough level of detail. It is worth mentioning that fractal dimension theory has also been applied in several scientific fields including the study of dynamical and mechanical systems [5, 6], diagnosis of diseases (such as osteoporosis [7] or cancer [8]), ecology [9], earthquakes [10], detection of eyes in human face images [11], analysis of the human retina [12], and brain computer interface systems [13], to name a few.
Usually, they are used both the Hausdorff and box dimensions, which can be defined on any metric space. Thus, while the former is “better” from a theoretical approach (since its definition is based on a measure), the latter is “better” from the viewpoint of applications, since it becomes easier to be empirically calculated or estimated. In this way, we should mention here that most empirical applications involving fractal dimension have been carried out in the context of Euclidean spaces through the box dimension.
The idea consisting of defining measures by means of coverings of certain subsets was first introduced by Carathéodory (c.f. [14]). Afterwards, Hausdorff applied this method to define the measures that now bear his name and showed that the middle third Cantor set has positive and finite measure of dimension equal to log2/log3 (c.f. [15]). Some properties and technical aspects regarding Hausdorff measures and dimensions have been developed mainly by Besicovitch [16], Besicovitch and his pupils [17], Falconer [2, 18], Feder [3], and Rogers [19].
On the other hand, it seems that the origins of box dimensions go back to the twenties, when they were first explored by pioneers of Hausdorff measure and dimension. Nevertheless, they were rejected for being less appropriate from a theoretical viewpoint. In this way, Bouligand adapted the Minkowski content to non-integral dimensions (c.f. [20]), and the classical definition of box dimension was provided by Pontrjagin and Schnirelman (c.f. [21]). Popularity of box dimension is mainly due to the possibility of its effective calculation and empirical estimation. Box dimension is also known as Kolmogorov entropy, entropy dimension, capacity dimension, metric dimension, information dimension, logarithmic density, . . . , etc.
The introduction of fractal structures, which were first sketched in [22] and then formally defined and applied in [23] to characterize non-Archimedeanly quasi-metrizable spaces, has allowed to formalize some topics on Fractal Geometry from both theoretical and applied viewpoints. A fractal structure is a countable collection of coverings of a given set which provides better approximations to it as deeper stages (called
Moreover, self-similar sets constitute a kind of fractals which can be always endowed with a fractal structure on a natural manner (first introduced in [25]). Along this survey, we shall provide some results allowing to calculate the fractal dimension of self-similar sets throughout an easy equation only involving the similarity ratios associated with the corresponding iterated function system.
First, we shall motivate each definition of fractal dimension and provide useful expressions to deal with its effective calculation. We collect some connections of each definition of fractal dimension with the classical definitions of fractal dimension, namely, both the box and the Hausdorff dimensions. In addition, we also provide some links to other fractal dimensions defined from a fractal dimension approach. Interestingly, we shall generalize the box dimension throughout the so-called fractal dimensions I, II, and III, whereas we shall generalize the Hausdorff dimension by means of fractal dimensions V and VI. It is also worth mentioning that fractal dimension IV constitutes a middle definition between Hausdorff and box dimensions.
Next, we summarize the content of each section in this survey.
In Section 2, we recall some concepts, results, and notations that become useful to develop a new theory of fractal dimension for fractal structures. This section is focused on the following topics: quasi-pseudometrics, fractal structures, iterated function systems, box and Hausdorff dimensions.
In both Sections 3 and 4, we formally introduce fractal dimension I and II models to calculate the fractal dimension of a set with respect to a fractal structure.They extend the classical box dimension theory to the more general context of fractal structures. Thus, if it is selected the so-called natural fractal structure which any Euclidean set can be always endowed with, then the box dimension remains a particular case. This idea allows us to consider a wide range of fractal structures to calculate the fractal dimension of any set. Unlike it happens with the classical theory of fractal dimension, the definitions we provide along this section can be computed in contexts where the box dimension can lack sense or cannot be calculated. In fact, the new models can be applied to calculate the fractal dimension of any space admitting a fractal structure as easy as the boxdimension in empirical applications. The results contained in this section were first contributed in [26].
In Section 5, it is provided a new model to calculate the fractal dimension of a set with respect to a fractal structure which generalizes the box dimension in Euclidean spaces. This has been carried out by means of a suitable discretization regarding both the Hausdorff measure and dimension. Thus, we shall provide some connections among this middle definition and the classical ones as well as with both fractal dimensions I and II explored in previous Sections 3 and 4. In this way, we shall generalize them and provide an easy expression to calculate the fractal dimension of strict self-similar sets not required to satisfy the so-called open set condition. The results appeared along this section were first contributed in [27].
In Section 6, we study how to generalize the Hausdorff dimension throughout three new models of fractal dimension for a fractal structure: two of them consist of discretizations of the Hausdorff dimension (fractal dimensions IV and V), while the remaining one becomes a new continuous approach to Hausdorff dimension from a fractal structure approach. We shall collect several results where the three new definitions are connected among them and also with fractal dimensions I, II, and III as well as with classical dimension. It is worth noting that the analytic construction of fractal dimension VI is based on a measure as it is the case of Hausdorff dimension. Additionally, we shall generalize Hausdorff dimension by means of fractal dimensions V and VI in the context of Euclidean sets endowed with their natural fractal structures. The results appeared in this section first appeared in [28].
The main goal in this section is to recall some notations, definitions, and notations that will result useful to tackle with a new theory of fractal dimension for fractal structures. In this way, we shall be focused on quasi-pseudometrics, fractal structures, iterated function systems, and Hausdorff and box dimension topics.
A quasi-pseudometric on a set
Moreover, if
3.
then
Let (
The concept of fractal structure was first introduced in [23] to characterize non-Archimedeanly quasi-metrizable spaces though it can be also used to deal with fractals. For instance, in [25] it was used to study attractors of iterated function systems.
Fractal structures constitute a powerful tool to introduce new models for a fractal dimension definition since it is a natural context where the concept of fractal dimension can be developed. Further, they will allow to calculate the fractal dimension in new spaces and situations.
A family Γ of subsets of a given space
Next, we provide a first approach to define a fractal structure on a set
(c.f. [23, Definition 3.1]) Let
A pre-fractal structure on Moreover, if Γ If
It is worth noting that covering Γ
(c.f. . [27, Remark 2.2]) To simplify the theory, the levels of any fractal structure
If
Self-similar sets are a kind of fractals that can be always endowed with a fractal structure in a natural way. Along this section, we recall a standard approach to construct attractors of iterated function systems. In addition, we shall also describe their natural fractal structure as self-similar sets. The results and properties described below are essential to understand some results appeared in upcoming sections.
Firstly, let
Next, let us recall the standard construction of self-similar sets provided by Hutchinson (c.f. [32]). Assume that (
It can be proved that 𝒲 is a contraction with respect to the Hausdorff metric
The next example describes analytically the so-called Sierpiński gasket, first defined in [33].
As it was mentioned previously, attractors of IFSs can be always endowed with a natural fractal structure, first sketched in [22] and formally defined later in [25]. It is worth pointing out that the latter provides the definition of a fractal structure as a mathematical concept, whereas the former deals with the (natural) fractal structure of a self-similar set. Next, we recall the description of such a fractal structure as provided in [25, Definition 4.4].
(c.f. [23, Definition 4.4]) Let
(c.f. [26, Remark 2.5]) Another suitable description concerning the levels of such a fractal structure is as follows: Γ1 = {
In Example 1, it was analytically described the IFS whose attractor is the Sierpiński gasket. Next, we describe the natural fractal structure that can be defined on this strict self-similar set.
(c.f. [26, Example 2]) The natural fractal structure on the Sierpiński gasket as a self-similar set is the countable family of coverings
Fractal dimension is one of the main tools applied to study fractals since it is a single quantity which throws useful information regarding their complexity when being examined with enough level of detail. It is worth noting that fractal dimension is usually understood as the classical box-counting dimension (box dimension along the sequel, for short) which is also known as information dimension, Kolmogorov entropy, capacity dimension, entropy dimension, metric dimension, logarithmic density, . . . , etc. (c.f. [2, Section 3.1]).
Though the Hausdorff dimension can be also considered a fractal dimension, in practical applications it is always used the box dimension since it can be easily calculated for a finite range of scales, which is the case of empirical applications.Popularity of the box dimension is mainly due to the possibility of its effective calculation and empirical estimation in Euclidean contexts. It is worth pointing out that, concerning applications of fractal dimension, the box dimension can be estimated as the slope of the regression line of a log−log graph plotted for a suitable discrete range of scales.
The basic theory on box dimension is covered in detail in [2]. Next, we recall the definition of the standard box dimension.
(c.f. [2, Section 3.1] and [2, Equivalent definitions 3.1]) The (lower/upper) box-counting dimension of a subset
where
The number of The number of The smallest number of sets of diameter at most The largest number of disjoint balls of radius
Notice also that the limit in Eq. (1) can be discretized by taking, for instance,
(c.f. [2, Section 3.1] and c.f. [26, Remark 2.7]) To calculate the (lower/upper) box dimension of any subset
The main purpose of this section is to include a sketch about the construction of both Hausdorff measure and dimension whose definitions and properties can be found out in [2, Chapter 2].
The first to define a measure by means of coverings of sets was Carathéodory in [14]. Later (1919), Hausdorff used this method to define the measures that now bear his name, and showed that the middle third Cantor set has positive and finite measure of dimension equal to log2/log3 [15]. A detailed study regarding the analytical properties of both Hausdorff measure and dimension was mainly developed by Besicovitch and his pupils.
The Hausdorff dimension, which is the oldest definition of fractal dimension, presents the best analytical properties. Indeed, note that this fractal dimension can be defined for any subset of a Euclidean (resp. metrizable) space and its definition is based on a measure which makes it quite appropriate from a mathematical viewpoint. Nevertheless, it presents some disadvantages, especially from the viewpoint of applications, since it can be hard to calculate or to estimate.
Thus, while this fractal dimension is “better” from a theoretical approach, the box-counting dimension is “better” for a wide range of applications. Next, let us recall the analytical construction of the Hausdorff dimension. Let (
Note that when
which is called the
It is worth mentioning that Hausdorff measure generalizes the classical Lebesgue measure for Euclidean subspaces. Indeed, if
From Eq. (2), it becomes clear that for any set
or equivalently,
In particular, if
Next theorem collects several properties that are satisfied by Hausdorff dimension as a dimension function. Such a result will be referred to afterwards to compare these properties for the different models of fractal dimension for fractal structures we shall introduce along upcoming sections.
(c.f. [2, Section 2.2])
It is also worth mentioning that the finite stability property consists of dim(
The countable stability property satisfied by the Hausdorff dimension (c.f. Eq. (5)) is the key for the next result.
There exists a Euclidean subset
It is also possible to calculate the Hausdorff dimension of a Euclidean subset
then we obtain the measure
The main goal in this section is to generalize the classical box dimension in the broader context of fractal structures. We state that whether the so-called
Let
(c.f. [26, Definition 3.1]) The natural fractal structure on every Euclidean space ℝ
Next, we highlight that natural fractal structures can be induced on Euclidean subsets.
(c.f. [26, Remark 3.2]) Natural fractal structures can be always defined on Euclidean subsets from previous Definition 4. For instance, the natural fractal structure (induced) on the closed unit interval [0,1] is defined as the family of coverings
The natural fractal structure on ℝ
Let us introduce our first generalized box dimension type model of fractal dimension.
(c.f. [26, Definition 3.3]) Let
The following remark becomes especially appropriate to deal with empirical applications involving the calculation of fractal dimensions.
(c.f. [26, Remark 3.4]) Fractal dimension I can be estimated in empirical applications throughout the slope of a regression line comparing level
The first theoretical result we provide in this section establishes that fractal dimension I generalizes classical box dimension in the context of Euclidean subsets endowed with their natural fractal structures.
Recall that Hausdorff dimension constitutes the main theoretical model of fractal dimension that we should be mirrored in when providing new definitions of fractal dimension. In this way, our next goal is to explore some theoretical properties from those listed in both Theorem 3 and Remark 4 for fractal dimension I.
Next step is to explore how box dimension and fractal dimension I are theoretically connected for any generalized-fractal space. To deal with, first we shall define the diameter of any level of a fractal structure and the diameter of any subset in a level of a fractal structure. Recall that a distance is a non-negative map
(c.f. [26, Definition 3.7]) Let
The diameter of level The diameter of where 𝒜
A feasible condition to be satisfied by a fractal structure consists of a geometric decrease regarding the sequence of diameters {
It is worth noting that Theorem 6 can be extended to deal with IFS-attractors. In fact, this is due to the fact that the sequence of diameters {
Next, we highlight how fractal dimension I depends on a selected fractal structure.
(c.f. [26, Remark 3.11]) There exists a Euclidean subset 𝒞 ⊂ ℝ endowed with two distinct fractal structures, say
A fractal structure is a kind of uniform structure. In fact, if there is no metric available in the space, the only way to “measure” a subset is by determining which level of the fractal structure contains that subset. In other words, it becomes quite natural that fractal dimension I depends on a fractal structure as well as box dimension depends on a metric.
Recall that fractal dimension I actually considers all the elements in level
(c.f. [26, Definition 4.1]) By a distance function (or a distance, for short), we shall understand a non-negative map
Diameters of subsets, coverings,
The second model for fractal dimension we shall provide with respect to a fractal structure is formulated in terms of a distance function.
(c.f. [26, Definition 4.2]) Let
where
In this subsection, we shall explore several conditions on the elements of each level of a fractal structure to reach the equality between fractal dimensions I and II. To tackle with, first we shall define the concepts of a semimetric on a topological space (c.f. [34, Definition 9.5]) and a semimetric associated with a starbase fractal structure.
(c.f. [34, Definition 9.5]) A semimetric on a topological space (a) (b) (c) The family { (c.f. [25, Theorem 6.5]) Let
It is worth pointing out that Eq. (6) implies
Moreover, it can be also proved that fractal dimension II generalizes both fractal dimension I and box dimension in the context of Euclidean subsets endowed with their natural fractal structures. That result, which extends former Theorem 4, is stated next.
Notice that Theorem 9 allows to calculate the box dimension of any plane subset by counting triangles instead of squares, for instance. To deal with, we could define a fractal structure on R2 consisting of triangulations whose triangles have a diameter of 1/2
In this section, we theoretically explore the behavior of fractal dimension II as a dimension function similarly to Proposition 5 for fractal dimension I.
In Remark 7, we highlighted the dependence of fractal dimension I on the selected fractal structure. Next, we point out the additional dependence of fractal dimension II on a metric. In particular, we shall justify why the fractal dimension II of the standard middle third Cantor set (endowed with its natural fractal structure as a self-similar set) equals its box dimension.
(c.f. [26, Remark 4.9]) Let 𝒞 denote the middle third Cantor set and
since level
Even more, though the value obtained in Remark 7 for
(c.f. [26, Remark 4.10]) Fractal dimension I only depends on a fractal structure whereas fractal dimension II also depends on a distance.
provided that
In this subsection, we provide an upper bound for both the Hausdorff and the box dimensions of any subset
Theorem 11 also allows to achieve an upper bound for the box dimension of IFS-attractors throughout their fractal dimension II values. That result is stated next.
Our next goal is to explore which properties underlying the natural fractal structure on any Euclidean space (c.f. Definition 4) could allow to generalize Theorem 9. With this aim, observe that given a scale
Let
It is worth pointing out that the main hypothesis in Theorem 13 to reach the equality between fractal dimension II and box dimension is necessary as the following counterexample highlights.
In addition, let
On the other hand, there are 8
for all
We also provide lower bounds for the ratios between
Accordingly, each subset
Let us recall when two sequences of positive real numbers are said to be of the same order.
(c.f. [26, Section 4]) Let
Thus, if it is assumed that all the elements in each family 𝒜
Fractal dimension II provides an upper bound concerning the box dimension of any Euclidean IFS-attractor (c.f. Corollary 12). Going beyond, it is even possible to reach that equality under certain conditions on the corresponding IFS. More specifically, this kind of result stands provided that the elements in each level of the fractal structure do not overlap “too much”. Hence, due to the shape of the elements in the natural fractal structure which any IFS-attractor can be endowed with, this restriction will rely on the similarities of the IFS. In this context, the so-called
Let ℱ be an IFS and 𝒦 be its IFS-attractor.
(c.f. [32, Section 5.2]) We understand that ℱ is under the OSC if there exists a (non-empty) bounded open subset 𝒱 ⊆ (c.f. [35]) Additionally, if 𝒱 ∩ 𝒦 ≠ ∅, then ℱ is said to satisfy the strong open set condition (SOSC).
Schief proved that both the OSC and the SOSC are equivalent for Euclidean IFSs (c.f. [36, Theorem 2.2]). In 1946, P.A.P. Moran contributed a strong result allowing the calculation of both the box and the Hausdorff dimensions for a certain class of Euclidean IFS-attractors throughout the (unique) solution of an equation only involving the similarity ratios associated with each similarity of the IFS (c.f. [37, Theorem III] and [2, Theorem 9.3]). That classical result is stated next.
(Moran, 1946).
Accordingly, under the OSC, the box dimension of any IFS-attractor equals its Hausdorff dimension, and that common value can be easily calculated from Eq. (7). The next result we provide guarantees the equality between the box dimension and the fractal dimension II of IFS-attractors lying under the OSC. Indeed, the calculation of these dimensions follows immediately from the number of similarities in the IFS and their common similarity ratio, as in [37, Theorem II].
We would like to point out that the hypothesis consisting of equal similarity ratios in Theorem 17 is necessary. Recall that Counterexample 14 implies that all the contractions involved in Theorem 17 must be similarities. Further, the following counterexample justifies why all the similarity ratios must be equal.
It is clear that their associated contraction ratios are
In both Sections 3 and 4, two novel definitions of fractal dimension for a fractal structure have been explored. Recall that fractal dimension I allows a selection involving a larger collection of fractal structures than box dimension. In fact, the natural fractal structure on any Euclidean subset (c.f. Definition 4) throws the classical box dimension as a particular case (c.f. Theorem 4).
On the other hand, though the fractal dimension II model allows the possibility that different diameter sets could appear in a level of a fractal structure, it does not actually distinguish among different diameter sets (c.f. Remark 9). Recall that we have to count the number of elements in each level of a fractal structure that intersect a given set
Let
as well as its asymptotic behavior,
Recall that Hausdorff dimension is fully determined throughout the (unique) value of
Observe that Eq. (10) is equivalent to
Letting
Hence, if ℋ
provided that
Hence, it becomes clear that
which implies inf{
However, unlike it happens with the
Interestingly, the problem consisting of the existence of the limit in Eq. (9) can be avoided whether the families
It is noteworthy that whether the families
Next, we provide the key definition of fractal dimension for a fractal structure we shall explore along this section.
(c.f. [27, Definition 4.2]) Let
where
and
The fractal dimension III of
It is worth pointing out that the sequence
(c.f. [27, Remark 4.3]) The following expressions are equivalent to calculate
inf{å inf{å
From Definition 13 of fractal dimension, it follows that the quantity
provided that
Additionally, the next remark becomes quite useful for fractal dimension III calculation purposes, since it highlights that it is no longer necessary to consider lower/upper limits to define
(c.f. [27, Remark 4.4]) Since
Along this subsection, we contribute several results to theoretically connect fractal dimension III with the classical definitions of fractal dimension, namely, both Hausdorff and box dimensions, as well as with fractal dimension II, previously explored in Section 4 for fractal structures.
The next corollary stands immediately from Theorem 20.
For a given subset
As well as the Hausdorff dimension definition is based on the
To deal with, let 𝒫(
It assigns the value 0 to the empty set, namely, It is monotonic increasing, i.e., if It is
for all sequence {
Notice that
(c.f. Eq. (12) or one of its equivalent expressions provided in Remark 10). Then
It is worth mentioning that though the two set function
Another issue naturally arising consists of determining some reasonable conditions on the elements in each level of a fractal structure to guarantee the equality among fractal dimension III and fractal dimensions I and
II. In this way, the following result we provide allows the calculation of fractal dimension III from the fractal dimension I formula provided that fractal structures having and appropriate size (of 1/2
Regarding the existence of the fractal dimension I of
(c.f. [27, Remark 4.11]) Under the hypothesis of Theorem 25, suppose that fractal dimension I does not exist for a given subset
Next step is to find out appropriate conditions regarding the size of the elements in each level of a fractal structure to reach the equality between fractal dimensions II and III.
Under the same hypothesis, a weaker result than Theorem 26 can be stated in the case that fractal dimension II does not exist. This is similar to Remark 12 allowing
(c.f. [27, Remark 4.13]) Under the hypothesis of Theorem 26, assume that fractal dimension II does not exist for a given subset
From both Remarks 12 and 13, we can state that fractal dimension III generalizes both fractal dimensions I and II for fractal structures having 1/2
It is worth mentioning that Corollary 27 allows the calculation of the fractal dimension III of any subset with respect to a fractal structure
Moreover, the following result establishes that all these fractal dimensions are equal in the context of Euclidean GF-spaces equipped with their natural fractal structures. In other words, fractal dimension III generalizes all the box dimension type models for fractal dimension including the classical one.
Previous Theorem 28 makes fractal dimension III to be understood as a hybrid approach to fractal dimension. In fact, though the analytical construction of fractal dimension III is based on a suitable discretization regarding the Hausdorff dimension, such a result states that fractal dimension III equals box dimension in the context of Euclidean subsets equipped with their natural fractal structures.
It is also worth pointing out that Theorem 28 also allows the calculation of fractal dimension III for Euclidean subsets throughout easier box dimension type expressions such as those provided in Sections 3 and 4.
Next, we collect several theoretical properties for fractal dimension III similarly to Proposition 5 for fractal dimension I and Proposition 10 for fractal dimension II.
Recall that in Theorem 13, some properties regarding the elements in each level of a fractal structure were provided to reach the equality between fractal dimension II and box dimension. It is worth pointing out that box dimension may be also defined for metrizable spaces. The next result we provide has been carried out in the spirit of Theorem 13 and generalizes Theorem 28.
As it was stated previously in Subsection 4.6, the issue concerning the calculation of the fractal dimension for IFS-attractors via algebraic expressions involving only a finite number of known quantities arises naturally for each new definition of fractal dimension. It is worth pointing out that this kind of theoretical results are inspired on classical Moran’s Theorem (c.f. Theorem 16) and usually assume that the similarities that give rise to the IFS-attractor are under the OSC hypothesis (c.f. Definition 12). In fact, recall that this constitutes the main constraint required to an IFS to reach the equality between the Hausdorff and the box dimensions of its strict self-similar set (c.f. Theorem 16).
The OSC is a
Next, we verify that Theorem 31 cannot be improved in the sense that the similarities
In addition, let
On the other hand, we affirm that
Therefore, each subset
As a consequence of Moran’s Theorem and Theorem 31, we have that the fractal dimension III of any IFS-attractor (endowed with its natural fractal structure) equals both its Hausdorff and box dimensions provided that the corresponding IFS is under the OSC.
Nevertheless, as the following counterexample highlights, the OSC hypothesis cannot be removed in previous Corollary 33.
Additionally, let
due to Theorem 31. Hence,
since all the elements in level
by Corollary 33, a contradiction.
Following Remark 34, we conclude that fractal dimension III does not coincide, in general, with Hausdorff dimension. Similarly, next we state that fractal dimension III may be also different from both fractal dimensions I and II.
(c.f. [27, Remark 4.24]) There exists an IFS
Further, let
On the other hand, notice that each covering Γ
by applying both Definition 8 and Theorem 15.
In this section, we study how to generalize the Hausdorff dimension throughout three new models of fractal dimension for a fractal structure: two of them will consist of an appropriate discretization regarding the Hausdorff dimension (fractal dimensions IV and V), whereas the remaining one will constitute a new continuous approach from a fractal structure viewpoint (fractal dimension VI). Several theoretical results to connect the three new definitions among them and also with fractal dimensions I, II, and III (introduced in previous sections) as well as with the classical definitions of fractal dimension will be provided. Moreover, we shall explore how the analytic construction regarding fractal dimension VI is based on a measure as it is the case of Hausdorff dimension. The main result in this section consists of a generalization of classical Hausdorff dimension in the context of Euclidean subspaces (endowed with their natural fractal structures) throughout both fractal dimensions V and VI. Finally, we shall contribute a result for IFS-attractors allowig to calculate these fractal dimensions via an easy equation only involving the similarity factors associated with the corresponding IFS.
Let
On the other hand, recall that the calculation of
Thus, let
If
Similarly to Counterexample 19, the following result points out that the condition
Since
In fact,
If (a) Assume that the covering of i. If that covering is {[0,1]}, then ii. Assume that the covering of (b) On the other hand, assume that the covering of On the other hand, if Accordingly, a new covering instead. In fact, for each element of the form Thus, å
Similarly to both Eqs. (14) and (15), the following expressions lead to a discrete fractal dimension for finite coverings, which will become especially appropriate to deal with empirical applications of fractal dimension [41]. Let us define
as well as
Thus, the asymptotic behavior of Eq. (17) plays a similar role to Hausdorff measure. Let
The following remark is analogous to Counterexample 35.
Another fractal dimension model described in terms of fractal structures can be sketched in the following terms. Let (
as well as the expression that follows:
The asymptotic behavior of
Let
where the sums are considered on
Hence,
Therefore, if
The previous models are formalized along the upcoming section.
The fractal dimension definitions that we shall explore along this section are provided next.
(c.f. [28, Definition 3.2]) Let
where
and define
for
In Definition 14 as well as in the next one, we shall assume that inf ⌀ = ∞. For instance, if
(c.f. [28, Definition 3.3]) Let
where
and define
The fractal dimension VI of F is given by the following critical point:
Equivalently, from both Definitions 14 and 15, it holds that
for
The next remark becomes especially useful, since it is not required to consider lower/upper limits (unlike it happens with box dimension) for
Since Since
Recall that Hausdorff dimension constitutes the key reference for new definitions of fractal dimension to be mirrored in. In fact, Hausdorff dimension satisfies some desirable properties as a dimension function which can be found out in both Theorem 3 and Remark 4. Similarly to Propositions 5 (for fractal dimension I), Proposition 10 (for fractal dimension II), and Proposition 29 (for fractal dimension III), next we shall explore the behavior of fractal dimensions IV-VI as dimension functions throughout a pair of theoretical results. The first of them collects some analytical properties which stand for fractal dimension IV.
An analogous result to Proposition 37 stands for both fractal dimensions V and VI.
Table 1 summarizes the behavior of all the fractal dimensions explored along this paper throughout the theoretical properties that each of them satisfy as dimension functions. It is worth mentioning that fractal dimensions V and VI behave more similarly to Hausdorff dimension than the other models, whereas fractal dimension I is the most similar definition to box dimension in this sense.
The table above summarizes all the theoretical properties satisfied by each definition of fractal dimension explored throughout this paper. We set 1 to denote that the corresponding property is satisfied by each fractal dimension and 0 otherwise (c.f. [42]).Theoretical properties dimB dimH Monotonicity 1 1 1 1 1 1 1 1 Finite stability 1 1 0 1 1 1 1 1 Countable stability 0 0 0 0 0 1 1 1 Countable sets 0 0 0 0 0 1 1 1 Closure dimension 0 1 1 1 0 1 1 1
In Section 6.2, we verified that fractal dimensions V and VI are quite close to classical Hausdorff dimension, at least from the viewpoint of the theoretical properties (as dimension functions) they satisfy. Following the above, the main goal in this section consists of going beyond so we can explore some conditions on a fractal structure to be able to connect fractal dimensions V and VI.
The first result we prove contains a first link between these fractal dimensions. It is worth mentioning that the only condition required therein concerns the sequence of diameters
Next, we provide a sufficient condition on the elements in each level of a fractal structure to guarantee the equality between fractal dimensions V and VI.
(c.f. [28, Definition 3.6]) Let
where
It is worth noting that wide families of fractal structures are diameter-positive. For instance, any finite fractal structure is diameter-positive. More specifically, next we provide several examples of wide families of fractal structures that are diameter-positive.
The following families of fractal structures are diameter-positive.
Any finite fractal structure. The natural fractal structure for any Euclidean subset (c.f. Definition 4). The natural fractal structure which any IFS-attractor can be always endowed with (c.f. Definition 2).
Under the diameter-positive condition for a fractal structure, it holds that fractal dimensions V and VI coincide.
Along this subsection, we shall connect fractal dimension V (and hence, fractal dimension VI, due to Lemma 39) with fractal dimension III, previously explored in Section 5. First of all, we state that the fractal dimension V is always
The result provided below gathers several connections among Hausdorff dimension (c.f. Subsection 2.5) and fractal dimensions for fractal structures: II (c.f. Definition 8), III (see Definition 13), IV and V (both of them described in Definition 14), and VI (c.f. Definition 15).
Interestingly, an additional link between fractal dimensions III and IV can be stated in the context of finite fractal structures.
We shall conclude this subsection by the next result which involves all the fractal dimensions from II to VI.
Our next goal is to show that the analytical construction regarding fractal dimension VI is also based on a measure. Indeed, unlike the set functions
Let (
If
An outer measure
Since the Method I provided in Theorem 45 may fail to provide a measure for which the open sets are measurable, the so-called
Let
We shall refer
As an immediate consequence of Theorem 46, we have that
The main goal in this subsection is to explore some connections of fractal dimensions IV, V, and VI with classical Hausdorff dimension. With this aim, next, we provide one of the main results in this paper, where we shall guarantee that fractal dimension V generalizes Hausdorff dimension in the context of Euclidean subsets endowed with their natural fractal structures.
As a consequence of Theorem 48, we can state that fractal dimension VI also equals the Hausdorff dimension of Euclidean subsets endowed with their natural fractal structures.
Interestingly, fractal dimension IV also generalizes Hausdorff dimension in the context of compact Euclidean subsets as the following results highlights.
In summary, fractal dimension V generalizes Hausdorff dimension for Euclidean subsets endowed with their natural fractal structures (c.f. Theorem 48). In addition, fractal dimension IV throws an upper bound to Hausdorff dimension in the same context as a consequence of Corollary 42 (1). The equality between Hausdorff dimension and fractal dimension IV has been reached for compact Euclidean subsets (c.f. Theorem 50). Going beyond, it becomes possible to weaken the hypothesis regarding the compactness of
Interestingly, all the fractal dimension models we have explored along this paper can be theoretically connected among them in the context of Euclidean subsets endowed with their natural fractal structures.
In particular, the fractal dimension IV model introduced in Definition 14 results especially appropriate from a theoretical viewpoint, since it becomes a middle dimension between classical fractal dimension definitions, namely, both box dimension and Hausdorff dimension, as we shall highlight along the next remark.
To conclude this section, we would like to point out that fractal dimension IV can also be applied for computational purposes. In other words, it becomes possible to computationally approach the fractal dimension IV of a compact subset, which, by Theorem 51 equals its Hausdorff dimension. Therefore, this fractal dimension will lead us to computationally deal with the calculation of the Hausdorff dimension of compact Euclidean subsets.
Next, we provide a preliminary example about how to computationally calculate the Hausdorff dimension of the middle third Cantor set.
and so on. Thus, we have Γ1 = {
In this example, we shall apply the following algorithm for fractal dimension IV calculation purposes: given
Example 54 has been provided to illustrate how to apply fractal dimension IV to deal with the effective calculation of the Hausdorff dimension of a compact Euclidean subset. It is worth mentioning that in [41], it was provided the first-known overall algorithm to calculate the Hausdorff dimension.