
Fractal interpolation function (FIF) is a new method of constructing new data points within the range of a discrete set of known data points. Consider the iterated functional system defined through the functions Wn(x,y)=(anx+en,αn(x)y+ψn(x)), n=1,…,N. Then, we may define the generalized affine FIF f interpolating a given data set {(xn,yn)∈I×R,n=0,1,…,N}, where I=[x0,xN]. In this paper, we discuss the smoothness of the FIF f. We prove, in particular, that f is θ-hölder function whenever ψn are. Furthermore, we give the appropriate upper bound of the maximum range of FIF in this case.
Citation: Najmeddine Attia, Taoufik Moulahi, Rim Amami, Neji Saidi. Note on fractal interpolation function with variable parameters[J]. AIMS Mathematics, 2024, 9(2): 2584-2601. doi: 10.3934/math.2024127
[1] | Najmeddine Attia, Neji Saidi, Rim Amami, Rimah Amami . On the stability of Fractal interpolation functions with variable parameters. AIMS Mathematics, 2024, 9(2): 2908-2924. doi: 10.3934/math.2024143 |
[2] | Xuezai Pan, Minggang Wang . The uniformly continuous theorem of fractal interpolation surface function and its proof. AIMS Mathematics, 2024, 9(5): 10858-10868. doi: 10.3934/math.2024529 |
[3] | Najmeddine Attia, Rim Amami . On linear transformation of generalized affine fractal interpolation function. AIMS Mathematics, 2024, 9(7): 16848-16862. doi: 10.3934/math.2024817 |
[4] | Baoxing Zhang, Yunkun Zhang, Yuanyuan Xie . Generating irregular fractals based on iterated function systems. AIMS Mathematics, 2024, 9(5): 13346-13357. doi: 10.3934/math.2024651 |
[5] | Hasib Khan, Jehad Alzabut, Anwar Shah, Sina Etemad, Shahram Rezapour, Choonkil Park . A study on the fractal-fractional tobacco smoking model. AIMS Mathematics, 2022, 7(8): 13887-13909. doi: 10.3934/math.2022767 |
[6] | Tunçar Şahan, Yunus Atalan . Novel escape criteria for complex-valued hyperbolic functions through a fixed point iteration method. AIMS Mathematics, 2025, 10(1): 1529-1554. doi: 10.3934/math.2025071 |
[7] | Martin Do Pham . Fractal approximation of chaos game representations using recurrent iterated function systems. AIMS Mathematics, 2019, 5(6): 1824-1840. doi: 10.3934/math.2019.6.1824 |
[8] | Jin Xie, Xiaoyan Liu, Lei Zhu, Yuqing Ma, Ke Zhang . The C3 parametric eighth-degree interpolation spline function. AIMS Mathematics, 2023, 8(6): 14623-14632. doi: 10.3934/math.2023748 |
[9] | Ruhua Zhang, Wei Xiao . What is the variant of fractal dimension under addition of functions with same dimension and related discussions. AIMS Mathematics, 2024, 9(7): 19261-19275. doi: 10.3934/math.2024938 |
[10] | Cemil Tunç, Alireza Khalili Golmankhaneh . On stability of a class of second alpha-order fractal differential equations. AIMS Mathematics, 2020, 5(3): 2126-2142. doi: 10.3934/math.2020141 |
Fractal interpolation function (FIF) is a new method of constructing new data points within the range of a discrete set of known data points. Consider the iterated functional system defined through the functions Wn(x,y)=(anx+en,αn(x)y+ψn(x)), n=1,…,N. Then, we may define the generalized affine FIF f interpolating a given data set {(xn,yn)∈I×R,n=0,1,…,N}, where I=[x0,xN]. In this paper, we discuss the smoothness of the FIF f. We prove, in particular, that f is θ-hölder function whenever ψn are. Furthermore, we give the appropriate upper bound of the maximum range of FIF in this case.
In approximation theory, fractal interpolation is an alternative to classical interpolation used when studying irregular curves. The motivation to study fractal interpolation functions (FIFs in short) comes from the fact that most time series studied in practice often exhibit fluctuations or abrupt changes that fractal interpolants can intrinsically model. The results indicate that the use of fractal interpolation in many areas (financial applications for example) is promising. The concept of the FIF was first introduced by Barnsley [1] via an iterated functional system (IFS in short) on a compact subset of R, which fundamentally acts as the pivot to construct fractals. Since then, this theory has become a useful and powerful tool in applied science and engineering [2,3,4,5,6]. Moreover, various and important properties of FIF have been proved, including smoothness, stability, and disturbance error (see for instance [7,8,9,10,11,12]).
Specifically, IFS is a collection of a complete metric space (X,d) with a finite set of continuous mappings w1,w2…,wN, for N≥2. One can find that there exists a compact set G=⋃Nn=1wn(G) referred to an invariant set or an attractor to the IFS. Moreover, Hutchinson's idea gives that the invariant compact set G is fully determined by the IFS, and also G is the limit of a sequence of sets that can be built by the members of the IFS (see for instance [13,14,15,16,17,18,19] for some extension of Hutchinson's framework). Recently, many researchers have been working on some extensions of the IFS framework (generalized contractions or more general spaces…). Fixed point theory plays a significant and vital role in the existence of invariant sets in different types of IFSs. To this end, many researchers have studied the existence of FIFs by using different results related to the fixed point theory [9,10,20,21,22,23].
A function g, defined on a set I, is said to be a Hölder continuous function with exponent θ (or shortly θ–Hölder function) when g satisfies
|g(x)−g(y)|≤c|x−y|θ,∀x,y∈I, |
for some positive constants c and 0<θ≤1. This relation is called the Hölder condition and, when θ=1, the function g is said to be Lipschitz in I with Lipschitz constant c. Let Φ:R+⟶R+ be a function. Then g is said to be a Φ-Hölder function if
|g(y)−g(x)|≤Φ(|y−x|),∀x,y∈I. |
We denote HΦ(I) the class of all Φ-Hölder functions on I [24,25] : It is well known that the class of HΦ(I) is closed and convex with respect to the pointwise supremum and infimum [24]. We say that Φ satisfies the doubling condition if there exists ξ≥1, depends on Φ and called Φ-Hölder constant, such that
Φ(bx)≤ξbΦ(x)for b≥1andΦ(bx)≤ξΦ(x)for b<1. |
Also, we denote by HdΦ(I) the family of Φ-Hölder functions such that Φ satisfies the doubling condition. The most important class of functions satisfies the doubling condition on R+ are the increasing and subadditive function, that is, Φ(x+y)≤Φ(x)+ϕ(y), with Φ(0)=0.
For n∈J:={1,…,N}, let αn:I⟶R be a Lipschitz function and ψn:I⟶R be a continuous function. In this paper, we consider the generalized affine FIF defined by
{Ln(x)=anx+enFn(x,y)=αn(x)y+ψn(x),n∈J, | (1.1) |
where, the real positive numbers an and en are determined by condition (2.1) and such that conditions (2.2) and (2.3) hold. This system is extensively studied when the functions {αn}n are constants (they are called vertical scaling factors) [7,8,12,26,27,28,29]. Further, the IFS can be selected suitably so that the corresponding FIF shares the quality of smoothness or non-smoothness. This depends on the choice of the vertical scaling factors and the functions ψn [26,27,30]. Then, choosing the appropriate vertical scale factors and functions ψn remind fundamental and they can fit the real rough curve precisely. Consider the case where the vertical scaling factor parameters are constants, then Chen [28], Chand and Kapoor [26,27] studied the smoothness of a class of FIFs and the smoothness of coalescence hidden variable FIFs, respectively, using the techniques of operator approximation. Moreover, and in the case where ψn, n∈J, are Lipschitz functions defined on I Yang and Yu [30] investigated the smoothness of a class of FIFs with variable parameters using new techniques. In this paper, we consider more general cases by letting ψn∈HdΦ(I), n∈J. Our smoothness results are obtained by evaluating |f(x)−f(y)| for x,y∈I. To this end, we study in Section 3, the effect of the choice of the function ψn on the FIF denoted by f, when ψn∈HdΦ(I). More precisely, we will prove the following result.
Theorem 1.1. Let f be the FIF generated by the IFS (1.1) and assume that ψn∈HdΦ(I). Let ζ:=maxn‖ψn‖∞, a=minnan, α:=maxn‖αn‖∞, ξr=ξa−r and C:=maxnCn where Cn is the Lipschitz constant of αn. For a given x,y∈Ln1n2…nk(I), we have
|f(x)−f(y)|≤k∑r=1ξrαr−1Φ(|x−y|)+k∑r=2ζαr−2C1−a|x−y|(a1−r−1)+2αk‖f‖∞. |
Note that Theorem 1.1 considers only the case when x,y∈Ln1n2…nk(I) (see definition in Section 3). However, for any x,y∈I, there exists k0 and {anj}j, such that
k0+1∏j=1anj≤|x−y]≤k0∏j=1anj,nj∈J. |
It follows that x,y∈Ln1n2…nk0(I) or x and y are belong to the adjacent two intervals with common boundary point denoted by z and then |f(x)−f(y)|≤|f(x)−f(z)|+|f(z)−f(y)|. The previous calculation may gives the useful upper bound of |f(x)−f(y)| for all x,y∈I. Moreover, , the most widely Φ-Hölder functions are the θ-Hölder functions. To this end, let Φ0(x)=s|x|θ, for some positive real number s and θ∈]0,1]. As an application of Theorem 1.1, we obtain the following result.
Corollary 1.1. Let f be the FIF generated by the IFS (1.1) such that αn are constant parameters, ψn∈HdΦ0(I). Assume that α:=maxn|αn|<a=minnan then the function f is a θ-Hölder on I, that is, there exists a positive constant d′ such that
|f(x)−f(y)|≤d′|x−y|θ,x,y∈I. |
Let (X,d) be a complete metric space. We define H(X) to be the set of all nonempty complex subsets of X and g:X⟶X. The mapping g will said to be a contraction if there exists c∈[0,1) such that
d(g(x),g(y))≤cd(x,y),∀x,y∈X. |
We define, on the set H(X), the Hausdorff metric dH defined as
d(A,B)=supx∈Ainfy∈Bd(x,y) and d(B,A)=supx∈Binfy∈Ad(x,y),A,B∈H(X), |
where dH(A,B)=max{d(A,B),d(B,A)}. It is well known [31], that the space (H(X),dH) is complete, and compact whenever X is compact. Now, we consider the IFS I={X, wn n∈J}, where wn:X⟶X is a continuous mapping for n∈J, and the Hutchinson operator W as a selfmapping of H(X) by
W(A)=N⋃n=1wn(A),∀ A∈H(X). |
A set G∈H(X) is said to be an attractor of the IFS if it satisfies G=N⋃n=1wn(G) that is W(G)=G. In fact, the IFS admits always at least one attractor [1]. Moreover, if the IFS is hyperbolic, that is each wn is a contraction, then we can prove that the operator W is a contraction mapping on (H(X),dH) [1,31].
The FIF can be defined as an interpolant function such that its graph is a fractal, or also as fixed point of maps using the notion of IFS. More precisely, let I=[x0,xN] be a real compact interval and let Δ={(xn,yn)∈I×R n∈J0:={0,1,…,N}} be a set of data, where x0<x1<⋯<xN, yi∈[a,b], with −∞<a<b<∞. For n∈J, set In=[xn−1,xn] and let Ln:I⟶In be a contractive homeomorphism such that
Ln(x0)=xn−1,Ln(xN)=xn,|Ln(x)−Ln(x′)|≤l|x−x′|, ∀ x,x′∈I, | (2.1) |
for some 0≤l<1. We consider N continuous mappings Fn:K:=I×[a,b]⟶R satisfying
Fn(x0,y0)=yn−1,Fn(xN,yN)=yn, | (2.2) |
|Fn(x,y)−Fn(x,y′)|≤|rn||y−y′|,∀x∈I,y,y′∈[a,b], | (2.3) |
for some rn∈(−1,1),n∈J. Now, we define the mapping Wn:K⟶In×R, as
Wn(x,y)=(Ln(x),Fn(x,y)),∀(x,y)∈K,n∈J. |
It is well known that the IFS {K,Wn:n∈J} has a unique attractor G. Moreover G is the graph of continuous function f:I⟶R that passes through all interpolation points (xn,yn), n∈J. This function is called FIF corresponding to the points (xn,yn), n∈J. It is a self-affine function since each affine transformation Wn maps the entire graph of the function to its section within the corresponding interpolation interval [1].
Let G={g:I⟶R, such that g is continuous, g(x0)=x0 and g(xN)=xN}. Then, (G,ρ) is a complete metric space, where ρ is a metric defined by
ρ(g,h)=‖g−h‖∞=max{|g(x)−h(x)|:x∈I},∀g,h∈G. |
Therefore, Read-Bajraktarevic operator T, defined on (G,ρ) by
T(g(x))=Fn(L−1n(x),g(L−1n(x))),x∈In,n∈J |
is a contraction mapping. Indeed, using (2.3), we obtain
‖T(f)−T(g)‖≤α‖f−g‖∞, |
where α:=maxn|αn|. Hence T possesses a unique fixed point f on G and then the FIF is the unique function satisfying the following functional relation
f(x)=Fn(L−1n(x),f(L−1n(x))),∀x∈In,n∈J. | (2.4) |
The most widely studied FIFs are defined by the following system
{Ln(x)=anx+en,Fn(x,y)=αny+ψn(x),n∈J, |
where the real constants an and en are determined by condition (2.1), ψn are some continuous functions such that conditions (2.2) and (2.3) hold, αn∈(−1,1) are free parameters, called vertical scaling factors of the transformations Wn, and have an important consequences on the properties of the FIF. Indeed, if we consider the case of equally spaced interpolation points, we obtain smooth or non-smooth fractal function depending on the scaling factors choice. More precisely, we have the box dimension D of the graph of the FIF defined by [31]
D:=1+log(∑Nn=1|αn|)log(N). | (2.5) |
In particular, if α1=⋯=αN=α then D=2+logN|α|. Nevertheless, there are questions about optimal choice of the vertical scaling factors αn, n∈J, so that the obtained curves fit as closely as possible the real values. There are different ways to measure the quality of fit of the interpolation, for example one can use the normalized mean squared error [22] (see also [32,33]).
In Figures 1–4, we plot the FIF associated to the interpolation points
Δ={(0,9),(0.2,11),(0.4,15),(0.6,8),(0.8,12),(1.0,10)}. |
However, different vertical scaling factors are employed in each construction. As, we can see, we obtain different shape of graph of FIF even, here, the vertical scaling factors were carefully selected, so that the box-counting dimension of each graph is equal to D=1.3988. Hence, the self-similarity of the fractal interpolation curve depends on the choice of the vertical scaling factors. To this end, considering more general case by choosing a variable parameters (αn(x) instead of constant parameters αn) provide a wide variety of systems for different approximations problems [30]. In the present work, we consider the IFS, with variable parameters [30], defined by (1.1). In this case, the FIF will be called generalized affine FIF and denoted by fα where α:=(α1,α2,…,αN) (or simply by f if there is no ambiguity).
In this section we consider the generalized affine FIF generated by the IFS defined by (1.1) in Section 1 We will assume, for n∈J, that the functions ψn∈HdΦ(I) and αn:I⟶R are Lipschitz functions, with Lipschitz constant Cn, such that α:=maxn‖αn‖∞<1, where ‖αn‖∞:=sup{αn(x);x∈I,n∈J}. Now, for x∈I, let
{Ln1n2…nk(x):=Ln1∘Ln2∘⋯∘Lnk(x)Ln1n2…nk(I):=Ln1∘Ln2∘⋯∘Lnk(I), |
where nj∈J, k≥1, j∈{1,…,k}. We define also, for j=1,…,k−1, a shift operator σj by σj(n1n2…nk)=nj+1…nk and
Lσj(n1n2…nk)(x)=Lnj+1…nk(x),1≤j≤k−1, |
while Lσk(n1n2…nk)(x)=x. In this paper, we consider the following convention ∏0j=1Sj(x)=1 for any family of functions {Sj}j.
First, we will prove the next lemma which will be useful in the proof of Theorem 1.1.
Lemma 3.1. Let k≥1, for all x,y∈Ln1n2…nk(I), nj∈J and j=1,…,k, there exist ˉx,ˉy∈I such that
(1) x=(∏kj=1anj)ˉx+∑kr=1(∏r−1j=1anj)enr and y=(∏kj=1anj)ˉy+∑kr=1(∏r−1j=1anj)enr.
(2) Let l∈{1,…,k}, then
|Lσl(n1n2…nk)(ˉx)−Lσl(n1n2…nk)(ˉy)|=(l∏j=1a−1nj)|x−y| | (3.1) |
and, there exits a positive constant ξl==ξa−l, such that
|ψnl(Lσl(n1n2…nk)(ˉx))−ψnl(Lσl(n1n2…nk)(ˉy))|≤ξlΦ(|x−y|). | (3.2) |
Proof. Using a successive iteration and induction (see [30], Lemma 3.1, and [34], Lemma 1) we have, for all nj∈J, j=1,…,k,
Ln1n2…nk(x)=(k∏j=1anj)x+k∑r=1(r−1∏j=1anj)enr. | (3.3) |
Since for every x,y∈I there exist ˉx,ˉy∈I such that x=Ln1n2⋯nk(ˉx) and y=Ln1n2⋯nk(ˉy), the first assertion follows. Now, for l∈{1,…,k}, we have
Lσl(n1n2…nk)(ˉx)=(k−l∏j=1anl+j)ˉx+k−l∑r=1(r−1∏j=1anl+j)enl+r=(k−l∏j=1anl+j)(k∏j=1a−1nj)[x−k∑r=1(r−1∏j=1anj)enr]+k−l∑r=1(r−1∏j=1anl+j)enl+r=(l∏j=1a−1nj)[x−k∑r=1(r−1∏j=1anj)enr]+k−l∑r=1(r−1∏j=1anl+j)enl+r. |
Similarly, we have Lσl(n1n2…nk)(ˉy)=(∏lj=1a−1nj)[y−∑kr=1(∏r−1j=1anj)enr]+∑k−lr=1(∏r−1j=1anl+j)enl+r and, as a consequence, we get (3.1). In addition, since Φ satisfies the doubling condition, there exists a constant ξ such that
|ψnl(Lσl(n1n2…nk)(ˉx))−ψnl(Lσl(n1n2…nk)(ˉy))|≤Φ(|Lσl(n1n2…nk)(ˉx)−Lσl(n1n2…nk)(ˉy)|)≤Φ(l∏j=1a−1nj|x−y|)≤ξlΦ(|x−y|). |
Now, we will give te prove of the Theorem 1.1. For this, let x,y∈Ln1n2…nk(I). Then, by Lemma 3.1, there exist ˉx,ˉy∈I such that x=(∏kj=1anj)ˉx+∑kr=1(∏r−1j=1anj)enr and y=(∏kj=1anj)ˉy+∑kr=1(∏r−1j=1anj)enr. Moreover, using [30, Lemma 3.2], we get for r≥2,
|r−1∏l=1αnl(Lσl(n1n2…nk)(ˉx))−r−1∏l=1αnl(Lσl(n1n2…nk)(ˉy))|≤r−1∑l=1αr−2C|Lσl(n1n2…nk)(ˉx)−Lσl(n1n2…nk)(ˉy)|≤r−1∑l=1αr−2C(l∏j=1a−1nj)|x−y|≤αr−2C|x−y|r−1∑l=1a−l=αr−2C1−a|x−y|(a1−r−1). | (3.4) |
Now, since f is the FIF generated by the system (1.1), we obtain, using the successive iteration and induction,
f(x)=f(Ln1n2…nk(ˉx))=[k∏j=1αnj(Lσj(n1n2…nk)(ˉx))]f(ˉx)+k∑r=1[r−1∏j=1αnj(Lσj(n1n2…nk)(ˉx))]ψnr(Lσr(n1n2…nk)(ˉx)). | (3.5) |
As a consequence, we get
|f(x)−f(y)|≤k∑r=1|[r−1∏j=1αnj(Lσj(n1n2…nk)(ˉx))]ψnr(Lσr(n1n2…nk)(ˉx))−[r−1∏j=1αnj(Lσj(n1n2…nk)(ˉy))]ψnr(Lσr(n1n2…nk)(ˉy))|+|[k∏j=1αnj(Lσj(n1n2…nk)(ˉx))]f(ˉx)−[k∏j=1αnj(Lσj(n1n2…nk)(ˉy))]f(ˉy)|. |
Now, using (3.2) and (3.4), we obtain
|f(x)−f(y)|≤k∑r=1[|r−1∏j=1αnj(Lσj(n1n2…nk)(ˉx))||ψnr(Lσr(n1n2…nk)(ˉx))−ψnr(Lσr(n1n2…nk)(ˉy))|+|r−1∏j=1αnj(Lσj(n1n2…nk)(ˉx))−r−1∏j=1αnj(Lσj(n1n2…nk)(ˉy))|]ψnr(Lσr(n1n2…nk)(ˉy)+2αk‖f‖∞≤k∑r=1ξrαr−1Φ(|x−y|)+k∑r=2ζαr−2C1−a|x−y|(a1−r−1)+2αk‖f‖∞, |
when we have used (3.2) and (3.4).
Remark 3.1. Let x,y∈I and k0 be an integer such that
k0+1∏j=1anj≤|x−y]≤k0∏j=1anj,nj∈{1,…,k0}. | (3.6) |
In order to simplify, we may take C=α ( this is the case, for example, when αn are constant functions α, for all n∈J). Therefore, using Theorem 1.1, we have, for all k≥k0,
|f(x)−f(y)|≤k∑r=1ξrαr−1Φ(|x−y|)+k∑r=2ζαr−11−a|x−y|(a1−r−1)+2αk‖f‖∞. |
Moreover, we can choose k large enough so that
|f(x)−f(y)|≤k∑r=1ξrαr−1Φ(|x−y|)+k∑r=2ζαr−11−a|x−y|a1−r, |
that is, we may take Ξk=2αk‖f‖∞−k∑r=2ζαr−11−a|x−y|≤0. Indeed, let A1:=ak0+1≤|x−y| by (3.6) and then
Ξk=2αk‖f‖∞−k∑r=2ζαr−11−a|x−y|≤2αk‖f‖∞−A1k∑r=2ζαr−11−a≤2αk‖f‖∞−A1ζα(1−a)(1−α)(1−αk−1)≤αk[2‖f‖∞−A1ζ(1−a)(1−α)αk−1+A1ζ(1−a)(1−α))]. |
Therefore, we only have to take k such that
A1ζ(1−a)(1−α)αk−1≥2‖f‖∞+A1ζ(1−a)(1−α):=Ξ1 |
or
αk−1≤A1ζΞ1(1−a)(1−α). |
In particular, take Φ(x)=|x|θ for θ∈(0,1]. It follows, since we can choose ξ=1 and then ξr=a−r, that
|f(x)−f(y)|≤k∑r=1ξrαr−1|x−y|θ+k∑r=2ζαr−11−a|x−y|a1−r≤|x−y|θα∞∑r=1(αa)r+ζa(1−a)α|x−y|∞∑r=2(αa)r≤1a−α|x−y|θ+ζα(1−a)(a−α)|x−y|. |
Example 3.1. The nowhere differentiable Weierstrass function is given by
fϕλ,l(x)=∞∑k=0λkϕ(lkx),x∈R, | (3.7) |
where l≥2 be an integer, 1/l<λ<1 and ϕ:R⟶R is a Z-periodic real analytic function. This function displays self-similarity on different scales (see Figure 5) and it's graph exhibits fractal-like behavior, with intricate and complex structure on all scales [35,36].
In this example, we consider the classical Weierstrass function f, that is, when ϕ(x)=cos(2πx). Let I=[0,1], N=l=3 and λ=1/2. Now consider
Wn(x,y)=(x+n−13,αn(x)y+ϕ(x+n−13)),(x,y)∈I×R, | (3.8) |
where αn(x)=12+(−1)1+⌊n/2⌋sin(2πx)4, for n∈{1,2,3}, where ⌊n/2⌋ means the integer part of x. In this case, we have Weierstrass function f is the FIF defined by {Wn}3n=1 [37]. Therefore, for n=1,2,3, we have ψn(x)=cos(2π(x+n−1)3) and then, for x,y∈I, we have
|ψn(x)−ψn(y)|=|sin(2π3x+2π3n−π6)−sin(2π3y+2π3n−π6)|≤2|sin(π3(x−y))|≤2π3|x−y|, |
where, we have used the inequality |sin(x)−sin(y)|≤sin(x−y2)cos(x+y2). Therefore, the function ψn is Φ-Hölder function with Φ(x)=2π3x (ξ=1). In this case, we have
a=13,ξr=a−r;C=14,α≤14andζ=1. |
Now, applying Remark 3.1, for k large enough and a given x,y∈Ln1n2…nk(I), we obtain
|f(x)−f(y)|≤k∑r=1ξrαr−1Φ(|x−y|)+k∑r=2ζαr−2C1−a|x−y|a1−r≤2π3α|x−y|k∑r=1(αa)r+ζC(1−a)α|x−y|k−1∑r=1(αa)r≤αa−α|x−y|(8π3+32)=(8π+92)|x−y|. |
In this section, we will prove some consequences of Theorem 1.1. For this, for each n∈J, let ψn∈HdΦ(I) and assume that ς:=∑∞r=1ξrαr<∞. We define the function
χ(x)=2M1(Φ(|x−y|)+|x−y|), |
where
M1=max{ςα,ςζCαξ(1−a)+2a‖f‖∞}. |
As a consequence of Remark 3.1, we obtain, the following result.
Proposition 3.1. Let f be the FIF generated by the IFS (1.1) such that ψn∈HdΦ(I) and αn are Lipschitz functions for each n∈J. Assume that ς:=∑∞r=1ξrαr<∞. Then f is χ-Hölder function on I.
Proof. Let x,y∈I, then there exists k0 such that (3.6) is satisfied. If k0=0 then we prescribe Ln1n2…nk0(I)=I. First, we consider the case when x,y∈Ln1n2…nk0(I), then, using the same notation as in Theorem 1.1, we have
|f(x)−f(y)|≤k0∑r=1ξrαr−1Φ(|x−y|)+k0∑r=2ζαr−2C1−a|x−y|a1−r+2αk0‖f‖∞≤k0∑r=1ξrαr−1Φ(|x−y|)+ζCαξ(1−a)|x−y|k0∑r=2ξr−1αr−1+2αk0ξk0+1ξ‖f‖∞|x−y|≤ςαΦ(|x−y|)+[ςζCαξ(1−a)+2a‖f‖∞]|x−y|,≤M1(Φ(|x−y|)+|x−y|)=12χ(|x−y|). |
where we have used the fact that ξk0αk0<1. Now, we consider the other case, that is, when x,y do not belong to the same subinterval Ln1n2…nk0(I) but (3.6) holds. Then, clearly the reals numbers x and y must belong to the adjacent two intervals with common boundary point denoted by z. It follows that
|f(x)−f(y)|≤|f(x)−f(z)|+|f(z)−f(y)|≤12χ(|x−z|)+12χ(|y−z|)≤χ(|x−y|). |
In the following we will give the proof of the Corollary 1.1. For this, let f be the FIF generated by the IFS (1.1) and let Φ(x):=Φ0(x)=s|x|θ, for some positive real number s and θ∈]0,1]. We assume that, for each n∈J, ψn∈HdΦ0(I). Again, we set ζ:=maxn‖ψn‖∞, a=minnan, α:=maxn|αn| such that α<a.
Now, under our hypothesis, we note that C=α, ξ=1, ξl=a−l and
ς=∞∑r=1(αa)r=αa−α<∞. |
Therefore, we deduce
|f(x)−f(y)|≤2ςαΦ(|x−y|)+2[ςζCα(1−a)+2a‖f‖∞]|x−y|≤2αsθa−α|x−y|θ+[2ζα3(a−α)(1−a)+4a‖f‖∞]|x−y|≤[2αsθa−α+2ζα3(a−α)(1−a)+4a‖f‖∞]|x−y|θ |
that is, the function f is a θ-Hölder on I.
Example 3.2. In this example, we consider the Weierstrass function defined in Example 3.1 by (3.7). Let I=[0,1] and the interpolating points x0=0<x1<⋯<xN=1 such that xn−xn−1=1/N (N=l). We consider the following system defined as
{Ln(x)=xN+n−1N,Fn(x,y)=αy+ϕ(x+n−1N), | (3.9) |
where α=λ. It is well known that the function f is a FIF [37]. Indeed, consider, for n∈J, the function
Wn(x,y)=(x+n−1N,αy+ϕ(x+n−1N)),(x,y)∈[0,1]×R. |
It follows that
f(Ln(x))=f(x+n−1N)=ϕ(x+n−1N)+α∞∑k=0αkϕ(Nkx)=ϕ(x+n−1N)+αf(x) |
and thus
Cf=N⋃n=1Wn(Cf). |
Therefore, for n∈J, we have ψn(x)=cos(2π(x+n−1)N). It follows, as in Example 3.1, that ψn is Φ-Hölder function with Φ(x)=2πNx and then we may choose ξ=1. In addition, choose α=12N, we get
a=1N,ξr=N−r,C=α=12Nandς=ζ=1. |
It follows, from Corollary 1.1, that
|f(x)−f(y)|≤[2αa−α2πN+2ζα3(a−α)(1−a)+4a‖f‖∞]|x−y|≤[4πN+12N(N−1)+4N‖f‖∞]|x−y|. |
Let I=[0,1], P={(nN,yn)∈R2,n∈J} be the interpolation points and D={nN∈I,n∈J0}. We define
L0(D)=D,L(D)=N⋃n=1Ln(D),and Lk(D)=L∘⋯∘L(D), |
k times composition. In this section, an interesting case of the system (1.1) will studied. Indeed, in [38], the author observed that we can use the theory of FIF to generate a family of continuous functions having fractal property from a given continuous function and with different values of fractal dimension. Let f∈C(I), the normed space of real valued endowed with the uniform norm continuous function on I, we define the following system
{Ln(x)=anx+enFn(x,y)=αn(x)y+f(Ln(x))−αn(x)b(x), | (4.1) |
where the real constants an and en are determined by condition (2.1), the functions αn:I⟶R are Lipschitz functions, with Lipschitz constant Cn such that α:=maxn‖αn‖∞<1 and b∈C(I) such that b(0)=f(0) and b(1)=f(1). Then the FIF generated by (4.1) will be denoted by fα which interpolates f at the nodes of the partition. Moreover, According to (2.4), the FIF fα satisfies the fixed point equation [30,38,39,40]
fα(x)=f(x)+αn(L−1n(x))(fα−b)(L−1n(x),for allx∈In,n∈J. | (4.2) |
Now, we will assume through this section that f and b are Φ1 and Φ2 Hölder functions with Hölder constants ξf and ξb respectively.
Lemma 4.1. Let fα be the FIF generated by the system (4.1) and assume that α=maxn‖αn‖∞<1. Then, there exists a positive constant A1 such that
|fα(x)−yn−1|≤αΓ1+ξf+αξb1−α,x∈In. |
Proof. We define, for k=1,2,…,
Γk=max{|fα(x)−y0|,x∈Lk−1(D)}andγk=maxn{|fα(x)−yn−1|,x∈Lk−1(D)∩In}. |
First, observe that
Γk≤maxn{|fα(x)−yn−1|,x∈Lk−1(D)∩In}+maxn{|yn−1−y0|}≤Γ1+γk. | (4.3) |
For x∈Lk(D)∩In, we have,
fα(x)=f(x)+αn(L−1n(x))(fα−b)(L−1n(x)) |
and then
|fα(x)−yn−1|≤|f(x)−f(n−1N)|+α|fα(L−1n(x))−y0|+α|b(L−1n(x))−y0|≤Φ1(|x−n−1N|)+αΓk−1+αΦ2(|L−1n(x)|)≤ξfΦ1(1)+αΓk−1+αξbΦ2(1)≤ξf+αΓk−1+αξb. |
We denote by A=ξf+αξb which nor depends on k. It follows, using (4.3), that
γk+1≤αΓk+A≤αγk+αΓ1+A≤β(αΓk−1+A)+αΓ1+A≤α2γk−1+α2Γ1+αΓ1+αA+A⋮≤k∑j=1αjΓ1+k−1∑j=0αjA≤αΓ1+A1−α. |
For any x∈In, there exits a sequence {xj}j∈In∩(⋃kLk(D)) such that xj⟶x and then limj→∞|fα(xj)−yn−1|=|fα(x)−yn−1|, by continuity of the function fα. Therefore, we get
|fα(x)−yn−1|≤αΓ1+ξf+αξb1−α,x∈In. |
Given a function S defined on I, we define the maximum range RS of S as
RS(I)=sups1,s2∈I|S(s1)−S(s2)|. |
Theorem 4.1. Let fα be the α-FIF the IFS (4.1) with interpolation points P. Assume that α<1, then
R˜fα(I)≤min{NαΓ1+Hf+αHb1−α,21−α(α‖b‖∞+‖f‖∞)}. |
Proof. From Lemma 4.1, we have
supIn|fα(x)−yn−1|≤αΓ1+Hf+αHb1−α. |
Now, let s1,s2∈I, then there exists n1≤n2∈J such that s1∈In1 and s2∈In2. It follows,
|fα(s1)−fα(s2)|≤|fα(s1)−yn1−1|+|yn1−1−yn1|+⋯+|yn2−1−fα(s2)|≤NαΓ1+Hf+αHb1−α. |
In the other hand, using (4.2), we obtain
R˜fα≤2‖fα‖∞≤2‖fα−f‖∞+2‖f‖∞≤2α1−α‖f−b‖∞+2‖f‖∞≤21−α(α‖b‖∞+‖f‖∞). |
as required.
Example 4.1. Let I=[0,1] and f(x)=x−x2. Observe that for any x,y∈I, we have
|f(x)−f(y)|≤|x−y|+|x2−y2|≤3|x−y| |
then the function f is Hölderian with exponent 1 and Hölder constant Hf=3. In this example, we consider the following perturbed system
{Ln(x)=xN+n−1NFn(x,y)=αy+f(Ln(x))−αb(x), | (4.4) |
where b(x)=f(x)/3. It follows that
‖fα−f‖∞≤α1−α‖f−b‖∞≤α6(1−α). |
In particular if α=1/6, we obtain
‖fα−f‖∞≤130. |
Therefore, we have
R˜fα(I)≤21−α(α‖b‖∞+‖f‖∞)≤11−α(α/12+1/4) |
and then R˜fα(I)=1960 for α=1/6.
In the present work, a class of generalized affine FIFs with variable parameters, where ordinate scaling is substituted by real-valued control function, is investigated. Their smoothness is discussed according to the choice of ψn, n∈J. We prove, in particular, that the FIF is θ-hölder function whenever ψn are. Our study is limited to functions ψn∈HdΦ(I) and it is worth studying more general cases, for example when doubling condition is not satisfied. Furthermore, we note that the thechnique using in this paper does not allows to study more general case, for example where Fn(x,y)=φn(y)+ψn(x) with φn are Matkowski contractions [22].
The authors declare they have not used Artificial Intelligence (AI) tools in the creation of this article.
The authors extend their appreciation to the Deanship of Scientific Research, Vice Pres- idency for Graduate Studies and Scientific Research at King Faisal University, Saudi Arabia, for financial support under the annual funding track [GRANT 5352].
The authors declare no conflit of interest.
[1] |
M. F. Barnsley, Fractal functions and interpolation, Constr. Approx., 2 (1986), 303–329. https://doi.org/10.1007/BF01893434 doi: 10.1007/BF01893434
![]() |
[2] |
G. E. Hardy, T. D. Rogers, A generalization of a fixed point theorem of Reich, Can. Math. Bull., 16 (1973), 201–206. https://doi.org/10.4153/CMB-1973-036-0 doi: 10.4153/CMB-1973-036-0
![]() |
[3] | M. A. Navascuès, M. V. Sebastian, Fitting curves by fractal interpolation: An application to the quantification of cognitive brain processes, In: Thinking in patterns, 2004,143–154. https://doi.org/10.1142/9789812702746_0011 |
[4] | P. R. Massopust, Fractal functions, fractal surfaces, and wavelets, Orlando: Academic Press, 1995. |
[5] |
A. Petrusel, I. A. Rus, M. A. Serban, Fixed points, fixed sets and iterated multifunction systems for nonself multivalued operators, Set-Valued Var. Anal., 23 (2015), 223–237. https://doi.org/10.1007/s11228-014-0291-6 doi: 10.1007/s11228-014-0291-6
![]() |
[6] |
X. Y. Wang, F. P. Li, A class of nonlinear iterated function system attractors, Nonlinear Anal. Theor., 70 (2009), 830–838. https://doi.org/10.1016/j.na.2008.01.013 doi: 10.1016/j.na.2008.01.013
![]() |
[7] |
M. A. Navascuès, M. V. Sebastian, Smooth fractal interpolation, J. Inequal. Appl., 2006 (2006), 78734. https://doi.org/10.1155/JIA/2006/78734 doi: 10.1155/JIA/2006/78734
![]() |
[8] |
A. K. B, Chand, G. P. Kapoor, Generalized cubic spline fractal interpolation functions, SIAM J. Numer. Anal., 44 (2006), 655–676. https://doi.org/10.1137/040611070 doi: 10.1137/040611070
![]() |
[9] |
J. Kim, H. Kim, H. Mun, Nonlinear fractal interpolation curves with function vertical scaling factors, Indian J. Pure Appl. Math., 51 (2020), 483–499. https://doi.org/10.1007/s13226-020-0412-x doi: 10.1007/s13226-020-0412-x
![]() |
[10] |
S. Ri, A new nonlinear fractal interpolation function, Fractals, 25 (2017), 1750063. https://doi.org/10.1142/S0218348X17500633 doi: 10.1142/S0218348X17500633
![]() |
[11] |
N. Vijender, Bernstein fractal trigonometric approximation, Acta Appl. Math., 159 (2018), 11–27. https://doi.org/10.1007/s10440-018-0182-1 doi: 10.1007/s10440-018-0182-1
![]() |
[12] |
P. Viswanathan, A. K. B. Chand, M. A. Navascuès, Fractal perturbation preserving fundamental shapes: Bounds on the scale factors, J. Math. Anal. Appl., 419 (2014), 804–817. https://doi.org/10.1016/j.jmaa.2014.05.019 doi: 10.1016/j.jmaa.2014.05.019
![]() |
[13] |
J. E. Hutchinson, Fractals and self-similarity, Indiana Univ. Math. J., 30 (1981), 713–747. https://doi.org/10.1512/iumj.1981.30.30055 doi: 10.1512/iumj.1981.30.30055
![]() |
[14] |
K. Lésniak, Infinite iterated function systems: A multivalued approach, Bulletin Polish Acad. Sci. Math., 52 (2004), 1–8. https://doi.org/10.4064/ba52-1-1 doi: 10.4064/ba52-1-1
![]() |
[15] | A. Mihail, R. Miculescu, The shift space for an infinite iterated function system, Math. Rep., 11 (2009), 21–32. |
[16] |
A. Mihail, R. Miculescu, Generalized IFSs on non-compact spaces, Fixed Point Theory Appl., 2010 (2010), 584215. https://doi.org/10.1155/2010/584215 doi: 10.1155/2010/584215
![]() |
[17] | N. A. Secelean, Countable iterated function systems, Far East J. Dyn. Syst., 3 (2001), 149–167. |
[18] |
F. Strobin, J. Swaczyna, On a certain generalization of the iterated function system, Bull. Aust. Math. Soc., 87 (2013), 37–54. https://doi.org/10.1017/S0004972712000500 doi: 10.1017/S0004972712000500
![]() |
[19] | K. R. Wicks, Fractals and hyperspaces, 1991. Berlin, Heidelberg: Springer. https://doi.org/10.1007/BFb0089156 |
[20] |
M. A. Navascués, C. Pacurar, V. Drakopoulos, Scale-free fractal interpolation, Fractal Fract., 6 (2022), 602. https://doi.org/10.3390/fractalfract6100602 doi: 10.3390/fractalfract6100602
![]() |
[21] |
S. Ri, New types of fractal interpolation surfaces, Chaos Solitons Fractals, 119 (2019), 291–297. https://doi.org/10.1016/j.chaos.2019.01.010 doi: 10.1016/j.chaos.2019.01.010
![]() |
[22] |
N. Attia, H. Jebali, On the construction of recurrent fractal interpolation functions using Geraghty contractions, Electron. Res. Arch., 31 (2023), 6866–6880. https://doi.org/10.3934/era.2023347 doi: 10.3934/era.2023347
![]() |
[23] |
N. Attia, M. Balegh, R. Amami, R. Amami, On the Fractal interpolation functions associated with Matkowski contractions, Electron. Res. Arch., 31 (2023), 4652–4668. https://doi.org/10.3934/era.2023238 doi: 10.3934/era.2023238
![]() |
[24] |
A. R. Goswami, Z. Páles, On approximately monotone and approximately Hölder functions, Period. Math. Hung., 81 (2020), 65–87. https://doi.org/10.1007/s10998-020-00351-0 doi: 10.1007/s10998-020-00351-0
![]() |
[25] | A. R. Goswami, Z. Páles, Characterization of approximately monotone and approximately Hölder functions, Math. Inequal. Appl., 24 (2021), 247–264. |
[26] | A. K. B. Chand, G. P. Kapoor, Smoothness analysis of coalescence hidden variable fractal interpolation functions, Int. J. Nonlinear Sci., 3 (2007), 15–26. |
[27] |
A. K. B. Chand, G. P. Kapoor, Stability of affine coalescence hidden variable fractal interpolation functions, Nonlinear Anal. Theor., 68 (2008), 3757–3770. https://doi.org/10.1016/j.na.2007.04.017 doi: 10.1016/j.na.2007.04.017
![]() |
[28] | C. Gang, The smoothness and dimension of fractal interpolation functions, Appl. Math. JCU, 11 (1996), 409–418. |
[29] |
Md. Nasim Akhtar, M. Guru Prem Prasad, M. A. Navascués, Box dimension of α-fractal function with variable scaling factors in subintervals, Chaos Solitons Fractals, 103 (2017), 440–449. https://doi.org/10.1016/j.chaos.2017.07.002 doi: 10.1016/j.chaos.2017.07.002
![]() |
[30] | H. Y. Wang, J. S. Yu, Fractal interpolation functions with variable parameters and their analytical proper ties, J. Approx. Theory, 175 (2013), 1–18. |
[31] | M. F. Barnsley, Fractals everywhere, Boston: Academic Press, 1988. |
[32] |
D. S. Mazel, M. H. Hayes, Using iterated function systems to model discrete sequences, IEEE Trans. Signal Process., 40 (1992), 1724–1734. https://doi.org/10.1109/78.143444 doi: 10.1109/78.143444
![]() |
[33] |
N. Vijender, V. Drakopoulos, On the Bernstein affine fractal interpolation curved lines and surfaces, Axioms, 9 (2020), 119. https://doi.org/10.3390/axioms9040119 doi: 10.3390/axioms9040119
![]() |
[34] |
H. Y. Wang, X. J. Li, Perturbation error analysis for fractal interpolation functions and their moments, Appl. Math. Lett., 21 (2008), 441–446. https://doi.org/10.1016/j.aml.2007.03.026 doi: 10.1016/j.aml.2007.03.026
![]() |
[35] |
H. J. Ren, W. X. Shen, A dichotomy for the Weierstrass-type functions, Invent. Math., 226 (2021), 1057–1100. https://doi.org/10.1007/s00222-021-01060-2 doi: 10.1007/s00222-021-01060-2
![]() |
[36] | T. Y. Hu, K. S. Lau, Fractal dimensions and singularities of the Weierstrass type functions, Trans. Amer. Math. Soc., 335 (1993), 649–665. |
[37] |
L. Jiang, H. J. Ruan, Box dimension of generalized affine fractal interpolation functions, J. Fractal Geom., 10 (2023), 279–302. https://doi.org/10.4171/JFG/136 doi: 10.4171/JFG/136
![]() |
[38] |
M. A. Navascués, Fractal polynomial interpolation, Z. Anal. Anwend., 24 (2005), 401–418. https://doi.org/10.4171/ZAA/1248 doi: 10.4171/ZAA/1248
![]() |
[39] | M. A. Navascués, Fractal trigonometric approximation, Electron. Trans. Numer. Anal., 20 (2005), 64–74. |
[40] | M. A. Navascués, Fractal functions on the sphere, J. Comput. Anal. Appl., 9 (2007), 257–270. |
1. | Najmeddine Attia, Rim Amami, On linear transformation of generalized affine fractal interpolation function, 2024, 9, 2473-6988, 16848, 10.3934/math.2024817 | |
2. | Najmeddine Attia, Neji Saidi, Rim Amami, Rimah Amami, On the stability of Fractal interpolation functions with variable parameters, 2024, 9, 2473-6988, 2908, 10.3934/math.2024143 |