Nonlinear Fractal Histopolation Function

Aiswarya T, Srijanani Anurag Prasad · 2025

Read Nonlinear Fractal Histopolation Function by Aiswarya T, Srijanani Anurag Prasad (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2509-18657.txt.

Introduction

A histogram sums data into bars, and a matching curve should keep the total area of those bars. Aiswarya T and Srijanani Anurag Prasad build such curves with fractal methods in this 2025 paper. Classic fractal interpolation gives smooth, continuous curves that pass through set points. Histopolation is different: the curve must preserve area, not height at points.

Their construction uses an iterated function system built on a generalized contraction idea. The result is a bounded curve that can be integrated, though it need not be continuous. Unlike older versions, their scaling factors may exceed one, which widens the shapes available. They derive the conditions under which the curve truly solves the histopolation problem. The paper extends the theory of fractal functions to a nonlinear setting. It is a proof-driven contribution aimed at approximation theory.

Catalog record

CREATORAiswarya T, Srijanani Anurag Prasad
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE28 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2509-18657

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Nonlinear Fractal Histopolation Function Aiswarya T, Srijanani Anurag Prasad Abstract A novel method for constructing a nonlinear fractal histopolation function associated with a given histogram is introduced in this paper. In contrast to classical fractal interpolation methods, which produce continuous and interpolatory functions, the proposed approach constructs a bounded, Riemann integrable function that is not necessarily continuous but preserves the area of a given histogram. An iterated function system based on Rakotch contractions- a generalisation of Banach contractions- is utilised, thereby extending the theoretical framework for fractal histopolation. Unlike previous formulations, the proposed construction of nonlinear fractal functions allows vertical scaling factors greater than one. The conditions for the nonlinear fractal function to be a solution for the histopolation problem are derived. Keywords : Histopolation, Iterated Function Systems, Rakotch Contraction, Fractal Interpolation Function MSC Classification : 28A80, 41A30, 37L30 1 Introduction Approximation theory is a branch of mathematics exploring methods to represent complex functions or data with more straightforward and manageable models. It aims to find efficient ways to approximate functions or datasets accurately using simpler mathematical constructs. Fractal approximation techniques provide a methodology to precisely approximate functions that exhibit self-similarity and a non-smooth nature. Interpolation is a commonly used method in approximation theory to find a continuous function whose graph coincides with a given data set. Histopolation is a method closely related to interpolation, but instead of approximating a function directly, histopolation preserves the area of the region below the function’s graph. Unlike interpolation, the histopolating function is not necessarily continuous. Fractal interpolation is a widely used approximation method with diverse practical applications. From image reconstruction to modelling of natural objects, fractal interpolation has proven its versatility. M.F. Barnsley introduced the fractal interpolation function(FIF) as a continuous function that exhibits self-similarity and interpolates a given data set barn1986 . The construction was based on the iterated function system(IFS) and the Banach contraction principle. Many researchers have worked on different aspects of fractal interpolation functions. Some of these studies constructed fractal interpolation functions of a more general nature, such as Hidden-variable FIFs barnsley89_2 , Hermite FIFs navas-sebas04 , Coalescence HFIFs chand07 , Super FIFs srijanani15 and Super CHFIFs srijanani21 . There have been studies conducted on properties like smoothness gang96 , regularity srijanani13_2 , approximation property navascues03 ; srijanani14_1 , multiresolution analysis hardin92 ; srijanani14_2 , reproducing kernel bouboulis11 ; srijanani19 , node insertion kocic03 ; srijanani13_1 , fractional calculus pan14 ; srijanani17 , and dimension property ruan09 . All the above constructions and studies had finite data points. Another extension from finiteness to countable data sets was introduced in secelean2003fractal . Using Rakotch contractions and Rakotch’s fixed point theorem, nonlinear FIFs were introduced for finite ri2018new and countable data sets pacurar2021countable . All the functions constructed in the works above were continuous and interpolatory. Recently, fractal histopolation functions were constructed by M.F. Barnsley and P. Viswanathan, which differ from FIFs as they are neither continuous (necessarily) nor interpolatory barnsley2023histopolating . The construction was based on the Banach contraction principle. This paper uses Rakotch’s fixed point theorem to construct nonlinear fractal histopolation functions. This construction is significant as it generalises the fractal histopolation function. Here, the IFS under consideration will consist of a more significant class of functions, the Rakotch contractions, rather than the small class of Banach contractions. A novel method for incorporating vertical scaling factors greater than one into the construction is presented, thereby enhancing its generality. The structure of the paper is as follows: Some definitions, notations and theorems are discussed in section 2 . A nonlinear fractal function is constructed in section 3 as the fixed point of a Read-Bajraktarević operator in the space of bounded functions on a closed and bounded interval. Conditions for the existence of the nonlinear fractal histopolation function as a solution for the given histopolation problem are found in section 4 . 2 Preliminary Facts This section contains notations, definitions and theorems necessary for the main result. Consider the metric space (X,\rho) . Definition 1 (Banach Contraction) . rudin1964principles Let g be a function on X . g is a Banach contraction if there exists a non-negative number k<1 such that \rho(g(x),g(y))\leq k\rho(x,y),\forall\ x,y\in X. The class of \psi -contractions generalises the concept of a Banach contraction on a metric space. Definition 2 ( \psi -Contraction) . strobin2015attractors Let \psi be a self map on \mathbb{R}_{+} . A function f on X is a \psi -contraction if \rho(f(t),f(s))\leq\psi(\rho(s,t)),\quad\forall\ s,t\in X. One of the widely discussed \psi -contractions is the Rakotch contraction. Definition 3 (Rakotch Contraction) . ri2018new Let f be a \psi -contraction. Suppose \psi satisfies \frac{\psi(s)}{s}<1 and \frac{\psi(t)}{t}\leq\frac{\psi(s)}{s},\forall\ t>s . Then, the function f is called a Rakotch contraction. The following examples illustrate that all Banach contractions are Rakotch contractions, but the converse is not true. Example 1 . Let (X,\rho) be a complete metric space and f:X\rightarrow X be a Banach contraction such that \rho(f(t),f(s))\leq k\rho(t,s),\forall~t,s\in X for some k\in[0,1) . Let \psi(t)=kt for the same k . Then, \rho(f(t),f(s))\leq\psi(\rho(t,s)) . Hence, f is a Rakotch contraction. Example 2 . Consider X=[0,1] with the Euclidean metric and the function h(t)=\frac{1}{1+t},\forall\ t\in X . \psi(t)=\frac{t}{1+t} . Then \forall\ t,s\in X , \displaystyle|h(t)-h(s)|=|\frac{1}{1+t}-\frac{1}{1+s}|=|\frac{s-t}{(1+t)(1+s)}|\leq\psi(|t-s|). Thus h is a Rakotch contraction. Note that the fixed point of h is \frac{\sqrt{5}-1}{2} . Let k\in[0,1) be any fixed constant. Let s be a positive real number such that s<\min\left\{-1+\frac{1}{k},\frac{\sqrt{5}-1}{2}\right\} . Then, |h(0)-h(s)|=|1-\frac{1}{1+s}|=|\frac{s}{1+s}|>ks=k|0-s|. Thus, h is not a Banach contraction. The existence of the fixed point of a Rakotch contraction is established in the following proposition. Proposition 1 . drakopoulos2021generalised Consider a Rakotch contraction f on a complete metric space (X,\rho) . Then, for any t\in X , \lim_{n\rightarrow\infty}f^{n}(t)=t_{0} and this point t_{0} is the unique fixed point of f. Definition 4 (Iterated Function System(IFS)) . barnsley2014fractals An iterated function system, \mathcal{I} , is defined as a complete metric space X together with a finite number of contraction maps, f_{1},f_{2},\ldots,f_{N} on X . This IFS is denoted as \mathcal{I}=\left\{X;f_{j},j=1,2,\ldots,N\right\} . Definition 5 (Attractor) . barnsley2014fractals A non-empty compact subset A of X is called an attractor for IFS \mathcal{I} if A=\bigcup_{j=1}^{N}f_{j}(A). The following proposition provides the existence of the attractor for an iterated function system consisting of Rakotch contractions. Proposition 2 . drakopoulos2021generalised Consider the IFS \left\{X;f_{1},f_{2},\dots,f_{N}\right\} on a complete metric space X where each of the f_{j} is a Rakotch contraction. Then, there exists a unique nonempty compact set B such that B=\bigcup\limits_{j=1}^{N}f_{j}(B). 3 Nonlinear Fractal Functions Let N>1 and \mathbb{N}_{N}=\left\{1,2,\ldots,N\right\} . Consider the interval I=[t_{0},t_{N}]\subset\mathbb{R} . Let \left\{t_{0},t_{1},\dots,t_{N}\right\}\subset I with t_{j-1}<t_{j} , for j\in\mathbb{N}_{N} . Let I_{j}=[t_{j-1},t_{j}) for j\in\mathbb{N}_{N-1} and I_{N}=[t_{N-1},t_{N}] . Let l_{j}:I\rightarrow I be the contractive maps l_{j}(t)=a_{j}t+b_{j} (3.1) satisfying l_{j}(t_{0})=t_{j-1} and l_{j}(t_{N})=t_{j} for all j\in\mathbb{N}_{N} . Set A=I\times\mathbb{R} . Define F_{j}:A\rightarrow\mathbb{R} as F_{j}(t,x)=c_{j}t+\delta_{j}s_{j}(x)+d_{j}, (3.2) where s_{j}(x) are Rakotch contractions with respect to the same function \psi for all j\in\mathbb{N}_{N} . The coefficients c_{j},d_{j} and \delta_{j} are constants where \delta_{j} are chosen as follows. If \sup\limits_{t>0}\frac{\psi(t)}{t}<1 , then there exists a \beta such that \sup\limits_{t>0}\frac{\psi(t)}{t}<\beta<1 . In this case choose \delta_{j} such that \delta=\max\limits_{j\in\mathbb{N}_{N}}|\delta_{j}|<\frac{1}{\beta} . If \sup\limits_{t>0}\frac{\psi(t)}{t}=1 , then choose \delta_{j} such that \delta=\max\limits_{j\in\mathbb{N}_{N}}|\delta_{j}|<1 . The functions F_{j} are bounded with respect to t . Remark 1 . Unlike the construction of

Read the complete text