Generalized escape criteria for fractals via convex viscosity approximation iterations Authors: Shuai Wang, Khaleel Ahmad, Yining Yang, Luminiţa-Ioana Cotîrlă, Daniel Breaz Published: 2026 ABSTRACT In this manuscript, we present the generation of Mandelbrot sets, Julia set fractals, and Biomorphs using a two-step viscosity approximation method applied to the complex function W(z)=zn+uz+r, where n≥2 and u,r∈C. The two-step viscosity approximation method is employed to establish new escape criteria for Julia sets, Mandelbrot sets, and Biomorphs. Subsequently, the viscosity approximation process is extended by incorporating m-convexity, and the corresponding escape criteria are generalized for these fractals. Further, we introduce the viscosity approximation process with s-convexity and develop the associated escape criteria for the same fractal structures. Additionally, we provide a comparative visualization of Mandelbrot sets, Julia set fractals, and Biomorphs using the standard viscosity approximation process, as well as its m-convex and s-convex variants, applied to the same complex polynomial function. High-resolution fractal images are produced using MATLAB R2024a with 50 iterations and a figure resolution of 800, highlighting the differences in the resulting structures for identical parameter values. 1. Introduction The captivating fractal structures that cover nature’s canvas reveal an amazing and deep intricacy that defies conventional geometry as patterns constantly reproduce themselves at smaller dimensions. Mandelbrot [1] established fractal geometry by showing how recursive mathematical functions might capture the irregularity of natural structures. Zhang et al. [2] suggested a chaos-based image encryption method employing Hilbert curves and H-fractals, so proving the significance of fractals in visual data security. Khalil et al. [3] demonstrated the computational efficiency of fractals in multimedia processing, so contributing to real-time imaging. Kumar [4] investigated the cryptographic possibilities of fractals. Ouyang et al. [5] created fractal drawings at the junction of mathematics and art that combined visual symmetry with recursive complexity. Krzysztofik [6] provided a complete review of fractals in antennas and metamaterials. Costanzo and Venneri [7] developed polarization-insensitive metamaterials for energy. Harris [8] underlined the way fractal design should be incorporated into architecture. Julia [9] was among the first to investigate the iteration of rational functions. Dang et al. [10] presented hypercomplex iterations and distance estimation for rendering high-dimensional fractals. Wang and Song [11] expanded this by producing Julia sets in bi-complex space and general Mandelbrot sets. Parise and Rochon [12] carried on this path using a study on tricomplex polynomial dynamics, so allowing analysis of fractals in three-complex-dimensional environments. Pickover [13] proposed the idea of biomorphs and generate the biologically inspired fractals. Kaboudian et al. [14] real-time interactive simulations of chaos and fractals. Jakubska-Busse et al. [15] investigated the effect of non-standard complex numbers in biomorph generation. In “Fractals Everywhere,” Barnsley [16] provided freely available, mathematically exact treatment of fractals. Devaney [17] matched this with providing an iterative function theory and chaotic dynamical system. Kumari et al. [18,19] proposed generalized multivalued iterated function systems and investigated their many structures in b-metric spaces. Hundertmark-Zaušková [20] showed by applying fixed-point iterations to moving geometries in fluid-structure interactions. Rahmani et al. [21] used fixed-point techniques for travel time estimate from sparse floating car data. Strogatz [22] reinforced the importance of convergence and iteration by offering a conceptual yet exact treatment of nonlinear dynamics and chaos. Usurelu et al. [23] combined numerical iteration with visual creativity. Mann [24] introduced the mean value iteration technique. Halpern [25] studied fixed points of non-expansive mappings. Nandal and Chugh [26] studied zero of accretive operators. Nandal et al. [27] investigated on variational inequalities involving pseudo-contractive operators. Postolache et al. [28] showed strong convergence results. Nandal et al. [29] proposed iterative schemes for maximal monotone operators. Zou et al. [30] investigated the Mandelbrot and Julia sets using Picard-Mann iteration. Shahid et al. [31] expanded this by including s-convexity. Chauhan et al. [32] used Ishikawa iteration and generate the Julia sets fractals. Saleem et al. [33] generalized the fractals in neutrosophic metric spaces. Kwun et al. [34] used Jungck-CR iteration with s-convexity and generated fractal. Kumari, and Chugh [35] introduced a new class of fractals by using the SP orbits and s-convexity. Jolaoso and Khan [36] proposed generalized escape time algorithms. Tassaddiq [37] used Jungck-Noor orbit and presented an extended escape criterion that improved control over fractal divergence patterns. Bhoria et. al [38] generated the Julia and Mandelbrot set fractals by using Picard Thakor iteration. Gdawiec, et al. [39] Generalized Logistic maps in the complex plane. Tassaddique et al. [40] studied fractal generation through fixed-point iteration processes that are modified. They were able to show that classical iteration schemes can be substantially enriched by changing them. The use of new iterative schemes has been also developed by Tanveer and Gdawiec [41], who applied the CR iteration scheme to create Mandelbrot sets related to the function zp+logct. Tassaddiq et al. [42] studied the Mann-iterative scheme in the framework of h-convexity of fractal generation. They introduced the application of convexity ideas to the iteration theory to give an analytical context that is generalized and guarantees convergence, and at the same time gives sets that are rich in fractal patterns. The escape conditions were also developed by Tassaddiq et al. [43] in the generation of fractals through the Picard Thakur hybrid iteration. Not only were their findings more efficient in computational efficiency. Tanveer, et. al [44] used Mann and Picard Mann iterations to analyze the Mandelbrot sets of zp+logct. They compared their study to show the influence of varying iteration schemes on the convergence rate, escape behavior and fractal geometry. The Hybrid Picard-S iteration scheme was presented in Tomar et al. [45] to conduct the detailed analysis of the escape of complex fractals. Generation of fractal using transcendental functions have also been extended by Alam et al. [46] who suggested a successful iterative process involving complex sine functions generate Julia and Mandelbrot sets. A new line of study connecting number theory with fractal visualization was introduced by Sharma et al. [47], who explored the application of the Fibonacci sequence in the development and visualization of fractals. Previous underlying literature by Tomar et al. [48] considered Julia and Mandelbrot sets that occurred when considering complex cosine functions by applying fixed-point iterative schemes. Alam et al. [49], who also developed escape conditions on orbits based on s-convexity and explored behavioral change in Mandelbrot and Julia sets. Their work presents a coherent mathematical structure of the analysis of divergence of orbits and transitions of stability in complex iterative systems. The dynamic of whole transcendental functions was studied by Prajapati et al. [50], who did a systematic study of Julia sets created using Mann-type schemes. Variants of Mandelbrot and Julia fractals of higher-order complex polynomial were studied by Tomar et al. [51]. Özdemir et. al [ 52–56] discussed the m-convexity, and s-convexity. ➢ In section 2, we discuss Julia set, Mandelbrot set, Mann iteration method, Halpern method, and viscosity approximation techniques, m-convexity, and s-convexity from existing literature. ➢ In section 2, we discuss Julia set, Mandelbrot set, Mann iteration method, Halpern method, and viscosity approximation techniques, m-convexity, and s-convexity from existing literature. ➢ In section 3, we discuss escape criteria Julia, Mandelbrot sets and Biomorph by using viscosity approximation process, viscosity approximation process with m-convexity, and viscosity approximation process with s-convexity. ➢ In section 3, we discuss escape criteria Julia, Mandelbrot sets and Biomorph by using viscosity approximation process, viscosity approximation process with m-convexity, and viscosity approximation process with s-convexity. ➢ In section 4, we show the algorithms (viscosity approximation method, viscosity approximation process with m-convexity, and viscosity approximation process with s-convexity for Julia set fractals and generate the Julia set fractals by using MATLAB R2024a. ➢ In section 4, we show the algorithms (viscosity approximation method, viscosity approximation process with m-convexity, and viscosity approximation process with s-convexity for Julia set fractals and generate the Julia set fractals by using MATLAB R2024a. ➢ In section 5, we discuss the algorithms for Mandelbrot set fractals via viscosity approximation method, viscosity approximation process with m-convexity, and viscosity approximation process with s-convexity. We generate the Mandelbrot set fractals by using MATLAB R2024a. ➢ In section 5, we discuss the algorithms for Mandelbrot set fractals via viscosity approximation method, viscosity approximation process with m-convexity, and viscosity approximation process with s-convexity. We generate the Mandelbrot set fractals by using MATLAB R2024a. ➢ In section 6, we demonstration the algorithms for Biomorph by using viscosity approximation method, viscosity approximation process with m-convexity, and viscosity approximation process with s-convexity. ➢ In section 6, we demonstration the algorithms for Biomorph by using viscosity approximation method, viscosity approximation process with m-convexity, and viscosity approximation process with s-convexity. ➢ In section 7, the ANI, Time plot, AET, and NAI are deliberated for viscosity approximation method, viscosity approximation process with m-convexity, and viscosity approximation process with s-convexity. ➢ In section 7, the ANI, Time plot, AET, and NAI are deliberated for viscosity approximation method, viscosity approximation process with m-convexity, and viscosity approximation process with s-convexity. ➢ In section 8, we discuss the comparison between viscosity approximation method, viscosity approximation process with m-convexity, and viscosity approximation process with s-convexity. ➢ In section 8, we discuss the comparison between viscosity approximation method, viscosity approximation process with m-convexity, and viscosity approximation process with s-convexity. ➢ In section 9, we derive conclusions from our results and future work. ➢ In section 9, we derive conclusions from our results and future work. The main objectives of this study are threefold. First, we establish new escape criteria for Julia sets, Mandelbrot sets, and Biomorphs generated by the complex polynomial W(z)=zn+uz+r, where n≥2 and u,r∈C using a two-step viscosity approximation method. Second, we extend the viscosity approximation process by incorporating m-convexity and s-convexity, and derive the corresponding escape criteria for these fractal structures. Third, we provide a comparative visualization and quantitative analysis (including Average Escape Time and Non-escaping Area Index) of the resulting fractals under the standard viscosity approximation and its m-convex and s-convex variants, using high-resolution images generated in MATLAB R2024a. Through this systematic investigation, we aim to demonstrate how the choice of convexity affects boundary sharpness, symmetry, complexity, and computational efficiency. 2. Preliminaries In this section, we discuss some fundamental definitions including Julia, Mandelbrot sets and Biomorph from existing literature. We denote C is a set of complex numbers and R is denoted for set of real numbers. Definition 2.1: [9] The filled Julia set Ftr of the complex function tr:C→C, where r∈C, is the set of points in the complex plan whose orbits under tr are bounded, i.e., where trj denotes the jth iteration of the function tr. The Julia set Jtr of tr is the boundary of the filled Julia set Ftr,, i.e., Jtr=∂Ftr. Benoit B. Mandelbrot [1], a French mathematician, helped Julia. Definition 2.2: [1] The Mandelbrot set Mtr of a function tr:C→C, is defined as the collection of all numbers r∈C for which the filled Julia set Ftr, remains connected, i.e., Definition 2.3: [24] Assume that D is a subset of a Banach space that is closed and convex. A self-mapping W is a non-expansive mapping in D. If a sequence {zn} satisfied the below pattern, then it is said to be Mann iteration process: where z0 is any starting point in D and a sequence αn in (0,1). Definition 2.4: [25] Suppose that {zn} is a sequence in C and z0∈C is an initial point. Then viscosity approximation method is defined as: where, g:C→C be a mapping and αn in (0,1). Definition 2.5: [51] Assume that I⊆R and a function f:[0,b]→R, where b∈I, is said to be m-convex if satisfied the following inequality, for x,y∈[0,b]: wherem, t∈[0,1]. Definition 2.6: [52] A function f:[0, ∞)→[0, ∞) is called first sense s-convex if it satisfied the below inequality, forx, y∈[0,∞): where t∈[0,1], ands∈(0,1]. Pickover [13] introduced the Biomorph in 1986. In which following modification was creating Julia set that are given below: where, R is denoted as escape radius. The Re(z) and Im(z) are real and Imaginary part of a complex number z∈C. 3. Escape criterion In this section, we will discuss the escape criterion for Julia set, Mandelbrot set, and Biomorphs for viscosity approximation method, viscosity approximation method with m-convexity, and viscosity approximation method with s-convexity. 3.1. Escape criterion for Viscosity Approximation method In this section, we discuss the escape criterion for Julia and Mandelbrot set fractals. We use α=infn∈Nαn where α∈(0,1) throughout the paper. We generalized the results [25], and generate Julia and Mandelbrot set fractals. The viscosity approximation process for any initial point z0∈C can be stated in the complex space as: for all j≥0, where α∈(0,1) and the functions W,g:C→C are self-mappings on C This two-step viscosity approximation iteration process has better control as compare to Picard or Mann iterations. The escape criterion is important in the visualization of Mandelbrot and Julia set fractals. we discuss an escape criterion for the following complex function that is defined as: where n≥2 and u,r∈C, by using viscosity approximation method given in (6). Let g(z)=az+b be a complex contraction with a,b∈C and |a|<1, we have −∣a∣>−1. Then by use −a≥−∣a∣, it follows that −a>−1, and hence−a−1<0. Theorem 3.1.1: Suppose |z0|≥max{|r|,|b|, (2+|u|)1n−1, (2+α+(1−α)|u|1−α)1n−1}, where α∈(0,1) and u,r,a,b∈C, and |a|<1. If the sequence {zj}N∪{0} satisfied the viscosity approximation iteration process that is given in (6). Then|zj|→∞ as j→∞. Proof: From (6), consider The assumption |z0|≥max{|r|,|b|, (2+|u|)1n−1, (2+α|a|+(1−α)|u|1−α)1n−1}, implies −|r|≥−|z0| and −|b|≥−|z0|, then we obtain We assume that |a|<1, we have The condition |z0|>(2+α+|u|(1−α)1−α)1n−1, so |y0|>|z0|. Therefore, the second step of viscosity approximation method is: Since, |z0|≥max{|r|,|b|, (2+|u|)1n−1, (2+α+(1−α)|u|1−α)1n−1}, implies −|r|≥−|z0|, and |y0|≥|z0|. We obtain, Since, |z0|≥max{|r|,|b|, (2+|u|)1n−1, (2+α+(1−α)|u|1−α)1n−1}, then|z1|≥|z0|. Thus, there exists a real number σ>0 such that |z1|>(σ+1)|z0|. Continuing this process we have Hence, |zj|→∞ as j→∞. We obtain the following corollary as a refinement of Theorem 3.1.1. Corollary 3.1.1: Assume that {zj} is a sequence defined in (6) and let for some k≥0. Then, there exists σ>0 such that |zk+j|>(1+σ)j|zk| and we have |zj|→∞ as j→∞. 3.2 Escape criterion for Viscosity Approximation method with m-convexity In this section, we generalized the result of section 3.1, with m-convexity. We discuss the escape criterion for Julia and Mandelbrot set fractals for viscosity approximation process with m-convexity where m∈[0, 1]. Definition 3.2.1: Assume that W,g:C→C are self-mappings on a complex set C. The viscosity approximation process with m-convexity is define as: for all j≥0, where α∈(0,1), andm∈[0, 1]. We discuss an escape criterion for the complex function that is defined (7) by using viscosity approximation method with m-convexity given in (8). Let g(z)=az+b be a complex contraction with a,b∈C and |a|<1. Theorem 3.2.1: Suppose |z0|≥max{|r|,|b|, (2+|u|)1n−1, (1+2α|m|(1−α)+|u|+1)1n−1}, where α∈(0,1) and u,r,a,b∈C, and |a|<1. If the sequence {zj}N∪{0} satisfied the viscosity approximation iteration process that is given in (8). Then|zj|→∞ as j→∞. Proof: The assumption |z0|≥max{|r|,|b|, (2+|u|)1n−1, (1+2α|m|(1−α)+|u|+1)1n−1}, implies −|r|≥−|z0| and−|b|≥−|z0|, and from (8), we have, We assume that |a|<1, so Since, |z0|≥max{|r|,|b|, (2+|u|)1n−1, (1+2α|m|(1−α)+|u|+1)1n−1}, so |y0|>|z0|. Therefore, the second step of viscosityapproximation method is: Since, |z0|≥max{|r|,|b|, (2+|u|)1n−1, (1+2α|m|(1−α)+|u|+1)1n−1}, implies −|r|≥−|z0|, and |y0|≥|z0|. We obtain, Since, |z0|≥max{|r|,|b|, (2+|u|)1n−1, (1+2α|m|(1−α)+|u|+1)1n−1}, then |z1|≥|z0|. Thus, there exists a real number σ>0such that |z1|>(σ+1)|z0|. Continuing this process we have Hence, |zj|→∞ as j→∞. We obtain the following corollary as a refinement of Theorem 3.2.1. Corollary 3.2.1: Assume that {zj} is a sequence define in (8) and let for some k≥0. Then, there exists σ>0 such that |zk+j|>(1+σ)j|zk| and we have |zj|→∞ as j→∞. Remark 3.2.1: In Theorem 3.2.1, if we take the value of m is 1. Then it same as Theorem 3.1.1. 3.3. Escape criterion for Viscosity Approximation method with s-convexity In this section, we generalized our results of section 3.1, with s-convexity. We discuss the escape criterion for Julia and Mandelbrot set fractals for viscosity approximation process with s-convexity where s∈(0, 1]. Definition 3.3.1: Assume that W,g:C→C are self-mappings on a complex set C. The viscosity approximation process with s-convexity is define as: for all j≥0, where α∈(0,1), ands∈(0, 1]. The escape criterion is important in the visualization of Mandelbrot and Julia set fractals. we discuss an escape criterion for the complex function that is defined in (7) by using viscosity approximation method with s-convexity given in (9). Let g(z)=az+b be a complex contraction with a,b∈C and |a|<1. Theorem 3.3.1: Suppose |z0|≥max{|r|,|b|, (2+|u|)1n−1, (2+sα(1−sα)+|u|)1n−1}, where α∈(0,1) and u,r,a,b∈C,and |a|<1. If the sequence {zj}N∪{0} satisfied the viscosity approximation iteration process that is given in (9). Then|zj|→∞ as j→∞. Proof: The assumption |z0|≥max{|r|,|b|, (2+|u|)1n−1, (2+sα(1−sα)+|u|)1n−1}, implies −|r|≥−|z0| and−|b|≥−|z0|, and from (9), we have, We assume that |a|<1, then we have Since, |z0|≥max{|r|,|b|, (2+|u|)1n−1, (2+sα(1−sα)+|u|)1n−1}, so |y0|>|z0|. Therefore, the second step of viscosity approximation method is: Since, |z0|≥max{|r|,|b|, (2+|u|)1n−1, (2+sα(1−sα)+|u|)1n−1}, implies −|r|≥−|z0|, and |y0|≥|z0|. We obtain, Since, |z0|≥max{|r|,|b|, (2+|u|)1n−1, (2+sα(1−sα)+|u|)1n−1}, then |z1|≥|z0|. Thus, there exists a real number σ>0 such that |z1|>(σ+1)|z0|. Continuing this process we have Hence, |zj|→∞ as j→∞. We obtain the following corollaries as a refinement of Theorem 3.3.1. Corollary 3.3.1: Suppose that {zj} is a sequence define in (6) and let for some k≥0. Then, there exists σ>0 such that |zk+j|>(1+σ)j|zk| and we have |zj|→∞ as j→∞. Remark 3.3.1: In Theorem 3.3.1, if we take the value of s is 1. Then it same as Theorem 3.1.1. 4. Algorithms for Julia set Fractals In this section, we discuss the algorithms of Julia sets fractals for complex polynomial W(z)=zn+uz+r where n≥2 and m,r∈C. Algorithm 1. Julia set generation for Viscosity approximation process Input: W(z)=zn+uz+r where n≥2 𝐚𝐧𝐝 u,r∈C. A ⊂ ℂ-area; K-the maximum number of iterations; α∈(0,1)-parameters for the viscosity approximation iterative method; g(z)=az+b, where a,b∈C and |a|<1; colour map [0,K]-color map with K+1 colors. Output: Julia set for areaA. i. 𝐑=𝐦𝐚𝐱{|r|,|b|,(2+|u|)1n−1,(2+α+(1−α)|u|1−α)1n−1}, ii. for z0∈A do iii. j=0 iv. 𝐰𝐡𝐢𝐥𝐞 |zj|