From Schreier graphs to secure s-boxes using a group-theoretic design framework Authors: Abdul Razaq, Muhammad Ramzan, Qin Xin, Hanan Alolaiyan, Dilshad Alghazzawi Published: 2026 ABSTRACT Secure substitution boxes (S-boxes) are a fundamental design element of modern symmetric cryptography because they provide the necessary nonlinearity to achieve confusion. This paper presents a new \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:8\times\:8$$\end{document} S-box construction framework, which combines group theory, graph theory and cryptography. The construction starts from the action of the modular transformation group on the projective line over \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{F}_{257}$$\end{document}, which produces Schreier graphs whose cycle structure is used to form an initial S-box. The design is then refined through a non-abelian permutation group acting on S-box positions. The proposed S-box is rigorously verified against conventional cryptographic criteria, including nonlinearity, differential uniformity, strict avalanche criterion, bit independence criterion, and linear approximation probability, which demonstrate that it is highly resistant to classical attacks. Besides, texture analysis is performed using the S-box in the image encryption setting that completely relies on substitution. The findings indicate enhanced randomness and reduced pixel correlation, as measured by gray-level co-occurrence matrix characteristics, such as contrast, correlation, energy, homogeneity, and entropy. The proposed methodology provides a strong theoretical foundation and demonstrates effectiveness for secure cryptographic and image processing applications. Introduction Background With the rapid growth of digital communication and data exchange, there is a need to safeguard sensitive and confidential data1–3. Cryptographic algorithms can be classified into two main categories: symmetric and asymmetric encryption algorithms. The most widely recognized symmetric encryption algorithms are the Advanced Encryption Standard (AES) and the Data Encryption Standard (DES)1. DES was created by IBM and became a U.S. federal standard in 1977. DES is now considered deprecated for modern cryptographic use because of its limited key length, although DES-based implementations may still be encountered in some legacy systems. Triple DES was later introduced as a transitional DES-based variant, but it is also no longer recommended for new encryption applications. The Advanced Encryption Standard (AES) was chosen by the National Institute of Standards and Technology (NIST) in 2001. Since then, AES has become widely used not only in the United States, but also globally. The key size of an encryption algorithm plays an important role in determining its security strength. AES is available in three different key lengths: 128 bits, 192 bits, and 256 bits. Claude Shannon2 proposed that a secure cryptosystem should primarily be based on two basic principles: confusion and diffusion. The S-box is a very important component of cryptographic algorithms, as it offers the required confusion property due to its nonlinear nature. In most symmetric ciphers such as AES and DES, it is the main element that causes this confusion. Being in many cases the only nonlinear component, it creates a complex relationship between plaintext and ciphertext. In one round of AES, four different operations are performed: SubBytes, ShiftRows, MixColumns, and AddRoundKey3–5. The S-box is applied in the SubBytes step, which is the only nonlinear transformation among these four. Cryptanalysis has shown that weaknesses in S-box design are a major vulnerability of DES. Therefore, the creation of secure cryptographic algorithms needs not only sound approaches to the design of S-boxes, but also sound criteria to assess them in terms of nonlinearity, algebraic degree, differential uniformity, and the strict avalanche criterion. This research paper develops a cryptographically secure 8 × 8 S-box and discusses its security properties. Literature review Khan and Shah6 suggested a new S-box design based on a fractional Rossler chaotic system that exhibited a high level of nonlinearity and was deemed appropriate for secure communication based on a thorough cryptographic study. Ahmad et al.7 proposed an effective 8 × 8 S-box architecture based on chaos theory and the Travelling Salesman Problem, which was shown to have better cryptographic measures than recent methods. Ullah et al.8 proposed a novel process of creating S-boxes using chaotic dynamics with a higher chaotic range. Their design had high cryptographic properties in various performance tests. Ahmed et al.9 proposed a cryptographic design technique based on an S-box design that integrates a discrete chaotic map and the firefly algorithm to optimize cryptographic performance, minimize a fitness function, and satisfy fundamental evaluation criteria such as nonlinearity and differential uniformity. The S-box generation technique presented in10 takes advantage of the activity of the modular group on the graph represented by adjacency matrices on finite fields, which leads to high-quality cryptographic performance and feasible efficiency in image encryption. Siddiqui et al.11 proposed an S-box design that used a modified Pascal triangle and elliptic curves and showed better nonlinearity and resistance to statistical attacks than more traditional S-boxes such as AES and Gray. Hayat et al.12 introduced a new dynamic S-box construction based on elliptic curves and total orderings over finite rings, which offered strong security properties with better performance than most current S-box generators. Fadhil et al.13 created secure S-boxes with good statistical characteristics, a fast inverse S-box, and the ability to be used in lightweight cryptographic systems using a 1D chaotic system. Another group-theoretic approach to building highly nonlinear S-boxes suggested by Razaq et al.14 demonstrated better cryptographic characteristics than the AES S-box and was effective in secure image encryption operations. Mahboob et al.15 combined finite field theory and permutation algorithms to design a secure S-box, the strength of which was validated by various statistical and cryptographic tests. Arshad16 proposed a new S-box design using elliptic curve points and group theory that showed great nonlinearity and was applicable in high-security image steganography. Abd-El-Atty17 suggested a quantum-inspired S-box based on quantum walks and PSO, which attained high entropy and security in image cryptosystems. Ullah et al.18 proposed a new S-box construction based on the fractional-order Rabinovich-Fabrikant chaotic system, improved with key-based permutation, with better statistical performance and effectiveness in secure image encryption applications. Tita et al.19 emphasised the use of S-boxes with low correlation values based on irreducible polynomials, which are appropriate for enhancing encryption against contemporary cyber threats. Liu et al.20 built secure keyed S-boxes in a systematic way through 2D chaotic maps and Boolean functions that were found to be more resistant to known cryptographic attacks and statistical weaknesses. The algorithm proposed by Shadab et al.21 provided a scalable and secure algorithm for S-box design, which was tested by comparing it to several cryptographic benchmarks. A secure S-box design with a 1D hybrid chaotic map and finite field theory was introduced by Jamal et al.22, and this design was found to have high nonlinearity and to remove the weaknesses of fixed points and short cycles. Bilal et al.23 designed a dynamic S-box based on Möbius transformations and elliptic curves, which is resistant to cryptanalytic attacks and can be used to perform encryption in real time. In24, finite algebraic structures are employed to construct Galois-field-based S-boxes and P-boxes for multiple color image encryption, enhancing confusion, diffusion, and resistance against common cryptographic attacks. Razaq et al.25 proposed a chaotic model that combines the Lorenz and Hénon systems to create powerful S-boxes with effective cryptographic properties. The hybrid chaotic map approach was utilized by Tolpa et al.26 to generate secure lightweight S-boxes to be integrated into low-resource cryptographic devices. Zhao et al.27 proposed an alternative encryption algorithm based on a complete chaotic map and dynamic S-boxes in order to achieve better image protection and anti-attack capabilities. Qayyum et al.28 proposed a finite-algebraic-structure-based multiple color image encryption scheme using Galois-field-driven S-boxes and P-boxes to improve confusion, diffusion, and resistance against common cryptographic attacks. The literature review shows that recent studies are dominated by chaos-based, optimization-based, and elliptic-curve-driven constructions, whereas comparatively fewer works develop S-boxes through an explicitly reproducible group- and graph-theoretic framework. To make this distinction clearer, Table 1 summarizes the main S-box construction approaches reported in the literature and highlights how the present work extends them through a Schreier-graph-based construction and a non-abelian positional refinement stage. Table 1Summary of previous S-box construction methods and the extension made in the present work.ReferencesPrevious workExtension made in the present workKhan and Shah6Designed an S-box using a fractional Rössler chaotic system with strong cryptographic properties.The present work replaces chaotic dependence with a reproducible group-theoretic construction based on Schreier graph cycles.Ahmad et al.7Proposed an 8 × 8 S-box using chaos theory and the Travelling Salesman Problem.Our method avoids heuristic path optimization and constructs the S-box through finite projective-line group action.Ullah et al.8Developed S-boxes through chaotic dynamics with an extended chaotic range.We provide an algebraic and graph-based construction that is more transparent and directly reproducible.Ahmed et al.9Combined a discrete chaotic map with the firefly algorithm to optimize S-box performance.The proposed framework generates candidates from modular group actions and improves them through a defined non-abelian permutation group.Xu and Feng10Used modular-group action on graph adjacency matrices over finite fields.We extend the modular-group idea by using Schreier graph cycle decomposition over (F_{257}\cup{\infty}) to form an 8 × 8 S-box.Siddiqui et al.11Constructed an S-box using a modified Pascal triangle and elliptic curves.Our work uses triangle-group action and Schreier graph orbits instead of elliptic-curve point arrangements.Hayat et al.12Introduced a dynamic S-box based on elliptic curves and total ordering over finite rings.The present study develops a cycle-ordering mechanism from Schreier graphs and further refines it by group-based positional permutations.Fadhil et al.13Designed secure S-boxes using a 1D chaotic system with good statistical features.Our method provides a non-chaotic construction with a clear algebraic origin and balanced cryptographic performance.Razaq et al.14Proposed a group-theoretic S-box with strong nonlinearity and image-encryption performance.We extend the group-theoretic direction by using modular transformation groups, Schreier graph cycles, and a non-abelian refinement stage.Mahboob et al.15Combined finite field theory and permutation algorithms for S-box design.We integrate finite-field action with graph-cycle interpretation and systematic group-based rearrangement.Arshad16Used elliptic-curve points and group theory for S-box generation.The present work uses triangle-group relations and Schreier graph structure rather than elliptic-curve point selection.Abd-El-Atty17Proposed a quantum-inspired S-box based on quantum walks and particle swarm optimization.Our framework offers a deterministic algebraic alternative without relying on quantum-inspired optimization.Ullah et al.18Designed an S-box using a fractional-order chaotic system with key-based permutation.We focus on the isolated strength of the S-box itself through substitution-only analysis and standard cryptographic tests.Tita et al.19Emphasized low-correlation S-boxes based on irreducible polynomials.Our approach adds Schreier graph structure and non-abelian positional refinement to obtain balanced security properties.Liu et al.20Built keyed S-boxes using 2D chaotic maps and Boolean functions.The present method gives a reproducible algebraic construction that does not depend on chaotic key generation.Shadab et al.21Presented a scalable S-box design tested against cryptographic benchmarks.We add an interpretable construction route in which the S-box entries arise from explicit group-action cycles.Jamal et al.22Used a hybrid chaotic map and finite field theory to construct an S-box.Our work uses finite-field structure together with modular transformation groups and Schreier graph decomposition.Bilal et al.23Designed dynamic S-boxes using Möbius transformations and elliptic curves.We employ Möbius-type transformations through triangle-group action and use the resulting cycles directly for S-box generation.Qayyum et al.24Used finite algebraic structures to construct Galois-field-based S-boxes and P-boxes for multiple color image encryption.Our work develops a Schreier-graph-based algebraic S-box construction with a clear and reproducible generation process.Razaq et al.25Combined Lorenz and Hénon chaotic systems to generate secure S-boxes.Our study presents a non-chaotic group- and graph-theoretic framework with explicit cycle construction.Tolpa et al.26Used hybrid chaotic maps to generate lightweight S-boxes.The present framework offers a transparent structural construction that may be adapted for lightweight cryptographic designs.Zhao et al.27Proposed image encryption using a complete chaotic map and dynamic S-boxes.We isolate the S-box effect through substitution-only image analysis before considering full encryption-pipeline effects.Qayyum et al.28Proposed a multiple color image encryption scheme using Galois-field-driven S-boxes and P-boxes.Our method focuses on direct S-box construction through group-action cycles and evaluates its isolated cryptographic strength. Table 1Summary of previous S-box construction methods and the extension made in the present work.ReferencesPrevious workExtension made in the present workKhan and Shah6Designed an S-box using a fractional Rössler chaotic system with strong cryptographic properties.The present work replaces chaotic dependence with a reproducible group-theoretic construction based on Schreier graph cycles.Ahmad et al.7Proposed an 8 × 8 S-box using chaos theory and the Travelling Salesman Problem.Our method avoids heuristic path optimization and constructs the S-box through finite projective-line group action.Ullah et al.8Developed S-boxes through chaotic dynamics with an extended chaotic range.We provide an algebraic and graph-based construction that is more transparent and directly reproducible.Ahmed et al.9Combined a discrete chaotic map with the firefly algorithm to optimize S-box performance.The proposed framework generates candidates from modular group actions and improves them through a defined non-abelian permutation group.Xu and Feng10Used modular-group action on graph adjacency matrices over finite fields.We extend the modular-group idea by using Schreier graph cycle decomposition over (F_{257}\cup{\infty}) to form an 8 × 8 S-box.Siddiqui et al.11Constructed an S-box using a modified Pascal triangle and elliptic curves.Our work uses triangle-group action and Schreier graph orbits instead of elliptic-curve point arrangements.Hayat et al.12Introduced a dynamic S-box based on elliptic curves and total ordering over finite rings.The present study develops a cycle-ordering mechanism from Schreier graphs and further refines it by group-based positional permutations.Fadhil et al.13Designed secure S-boxes using a 1D chaotic system with good statistical features.Our method provides a non-chaotic construction with a clear algebraic origin and balanced cryptographic performance.Razaq et al.14Proposed a group-theoretic S-box with strong nonlinearity and image-encryption performance.We extend the group-theoretic direction by using modular transformation groups, Schreier graph cycles, and a non-abelian refinement stage.Mahboob et al.15Combined finite field theory and permutation algorithms for S-box design.We integrate finite-field action with graph-cycle interpretation and systematic group-based rearrangement.Arshad16Used elliptic-curve points and group theory for S-box generation.The present work uses triangle-group relations and Schreier graph structure rather than elliptic-curve point selection.Abd-El-Atty17Proposed a quantum-inspired S-box based on quantum walks and particle swarm optimization.Our framework offers a deterministic algebraic alternative without relying on quantum-inspired optimization.Ullah et al.18Designed an S-box using a fractional-order chaotic system with key-based permutation.We focus on the isolated strength of the S-box itself through substitution-only analysis and standard cryptographic tests.Tita et al.19Emphasized low-correlation S-boxes based on irreducible polynomials.Our approach adds Schreier graph structure and non-abelian positional refinement to obtain balanced security properties.Liu et al.20Built keyed S-boxes using 2D chaotic maps and Boolean functions.The present method gives a reproducible algebraic construction that does not depend on chaotic key generation.Shadab et al.21Presented a scalable S-box design tested against cryptographic benchmarks.We add an interpretable construction route in which the S-box entries arise from explicit group-action cycles.Jamal et al.22Used a hybrid chaotic map and finite field theory to construct an S-box.Our work uses finite-field structure together with modular transformation groups and Schreier graph decomposition.Bilal et al.23Designed dynamic S-boxes using Möbius transformations and elliptic curves.We employ Möbius-type transformations through triangle-group action and use the resulting cycles directly for S-box generation.Qayyum et al.24Used finite algebraic structures to construct Galois-field-based S-boxes and P-boxes for multiple color image encryption.Our work develops a Schreier-graph-based algebraic S-box construction with a clear and reproducible generation process.Razaq et al.25Combined Lorenz and Hénon chaotic systems to generate secure S-boxes.Our study presents a non-chaotic group- and graph-theoretic framework with explicit cycle construction.Tolpa et al.26Used hybrid chaotic maps to generate lightweight S-boxes.The present framework offers a transparent structural construction that may be adapted for lightweight cryptographic designs.Zhao et al.27Proposed image encryption using a complete chaotic map and dynamic S-boxes.We isolate the S-box effect through substitution-only image analysis before considering full encryption-pipeline effects.Qayyum et al.28Proposed a multiple color image encryption scheme using Galois-field-driven S-boxes and P-boxes.Our method focuses on direct S-box construction through group-action cycles and evaluates its isolated cryptographic strength. Summary of previous S-box construction methods and the extension made in the present work. Motivation, research gap, and contributions Most of the current methods are based on randomness or heuristic optimization, which are not necessarily well-grounded in mathematics or easily reproducible. Moreover, group-theoretic and graph-based constructions, which are potentially fruitful, have not been exhaustively studied in a coherent framework that systematically utilizes their structural richness. Specifically, the potential of modular transformation groups and Schreier graphs to produce highly nonlinear and well-structured S-boxes is not yet well developed in the literature. Also, little focus has been placed on the assessment of S-box performance in terms of texture and image-processing characteristics, which is a crucial aspect in real-world encryption. In order to fill these gaps, this paper presents a new S-box construction model based on Schreier graph theory and modular transformation group actions. The main contributions of the work are as follows: (i) a mathematically rigorous approach for constructing permutation structures through the action of triangle groups on finite projective lines; (ii) the creation of an initial S-box based on cycle decomposition and its systematic improvement using a non-abelian permutation group; (iii) extensive testing using standard cryptographic metrics to demonstrate high resistance to known attacks; (iv) the use of texture-based analysis to confirm the usefulness of the proposed S-box in image encryption applications. Algebraic preliminaries This section presents the group-theoretic background, namely, the theory of the modular transformation group and its action, on which our S-box is constructed. The modular transformation group M is an infinite, non-abelian group generated by two bijective transformations \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\sigma\:$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\tau\:$$\end{document} acting on the projective line over a finite field, denoted by \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{F}_{p}\cup\:\{\infty\:$$\end{document}}, defined as \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\sigma\:\left(z\right)=\frac{-1}{z}$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\tau\:\left(z\right)=\frac{z-1}{z}$$\end{document}. All arithmetic in these transformations is performed in \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{F}_{p}$$\end{document}, that is, modulo p; division by a nonzero element means multiplication by its modular inverse, with the usual projective-line convention for the point \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\infty\:$$\end{document}. The orders of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\sigma\:$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\tau\:$$\end{document} are 2 and 3, respectively, leading to the group presentation \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:M = < \sigma \:,\:\tau \:\:|\:\sigma \:^{2} = \tau \:^{3} = 1 > .$$\end{document}. Although M is infinite, it can act on finite sets, producing finite Schreier graphs with rich combinatorial structure. Specifically, when\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\:M$$\end{document} acts on \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{F}_{p}\cup\:\{\infty\:$$\end{document}}, where p is prime, the resulting Schreier graph is composed of interconnected triangles. This triangular structure arises because the generator \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\tau\:$$\end{document} has order 3, thus forming 3-cycles in its permutation representation. For example, under this action,\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\sigma\:\left(0\right)=\infty\:,\:\:\mathrm{a}\mathrm{n}\mathrm{d}\:\tau\:\left(0\right)=1,\:\tau\:\left(1\right)=\infty\:,\:\tau\:\left(\infty\:\right)=0.$$\end{document} Hence, to make the action of these fractional linear transformations well-defined, we work on the projective line over \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{F}_{p}$$\end{document}, defined as the set \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{F}_{p}\cup\:\{\infty\:$$\end{document}}. Example 2.1 Let us consider the action of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:M\:$$\end{document}on \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{F}_{19}\cup\:\left\{\infty\:\right\}=\{\mathrm{0,1},2,\dots\:,18,\infty\:\}$$\end{document}. The permutations induced by \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\sigma\:$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\tau\:$$\end{document} on this set are as follows:\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\sigma\:=(0,\infty\:)\left(\mathrm{1,18}\right)\left(\mathrm{2,9}\right)\left(\mathrm{3,6}\right)\left(\mathrm{4,14}\right)\left(\mathrm{5,15}\right)\left(\mathrm{7,8}\right)\left(\mathrm{10,17}\right)\left(\mathrm{11,12}\right)\left(\mathrm{13,16}\right).$$\end{document}\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\tau\:=(0,\infty\:,1)\left(\mathrm{2,10,18}\right)(\mathrm{3,7},9)\left(\mathrm{4,15,6}\right)\left(\mathrm{5,16,14}\right)\left(\mathrm{11,13,17}\right)\left(8\right)\left(12\right).$$\end{document} In the coset diagram, each 3-cycle of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\tau\:$$\end{document} corresponds to a triangle, and each 2-cycle of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\sigma\:$$\end{document} (which is a transposition) represents an edge connecting two vertices. These triangles are connected to each other via the edges defined by \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\sigma\:$$\end{document}. The resulting coset diagram (shown in Fig. 1) therefore consists of six triangles linked according to these cycles. In this representation, every vertex is fixed by \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\sigma\:}^{2}\:$$\end{document}and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\tau\:}^{3},\:$$\end{document}visually encoding these group relations. The coset diagram is thus a picture of the group action. The parametrization method of Mushtaq29 allows us to construct diagrams where the composite transformation \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\sigma\:\tau\:$$\end{document} also has a finite order p, thereby satisfying the additional relation \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\left(\sigma\:\tau\:\right)}^{n}=1$$\end{document} and realizing the triangle group \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\left(\mathrm{2,3},n\right)$$\end{document}. Fig. 1Coset diagram for the action of M on \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{F}_{19}\cup\:\left\{\infty\:\right\}$$\end{document}. Prepared using microsoft power point 2021. URL: https://www.microsoft.com/en-us/microsoft-365/powerpoint. Coset diagram for the action of M on \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{F}_{19}\cup\:\left\{\infty\:\right\}$$\end{document}. Prepared using microsoft power point 2021. URL: https://www.microsoft.com/en-us/microsoft-365/powerpoint. Mushtaq29 introduced a parametrization method that enables the construction of a family of coset diagrams. For a chosen parameter \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\theta\:\in\:{F}_{p\:},\:$$\end{document}this method generates a graph whose associated group action satisfies the relations of a triangle group\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\left( {2,3,n} \right) = < \sigma \:,\tau \:\:|\sigma \:^{2} = \tau \:^{3} = \left( {\sigma \:\tau \:} \right)^{n} = 1 > ,$$\end{document} where the order n of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\sigma\:\tau\:\:$$\end{document}is determined by the value of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\theta\:$$\end{document}. To construct such graphs, one defines two general Möbius transformations:\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\sigma\:\left(z\right)=\frac{az+kc}{cz-a},\:\:\tau\:\left(z\right)=\frac{dz+kf}{fz-d-1}$$\end{document} For more details and a complete proof of this method, we refer to29. The parameters \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:a,c,d,f,k,r,s,\varDelta\:$$\end{document} are chosen to satisfy certain algebraic relations connecting \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\theta\:$$\end{document} with the order of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\sigma\:\tau\::$$\end{document}1\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\theta =\frac{{{r^2}}}{\Delta }$$\end{document}2\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${r^2}+k{s^2}=3\Delta ~~$$\end{document}3\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${d^2}+d+k{f^2}+1=0$$\end{document}4\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$a\left( {2d+1} \right)+2kcf - r=0$$\end{document}5\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$2af - c\left( {2d+1} \right) - s=0$$\end{document} Solving this system of equations for a given \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\theta\:$$\end{document} determines the exact forms of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\sigma\:\:$$\end{document}and\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\:\tau\:$$\end{document}, and hence the structure of the corresponding Schreier graph. The relationship between the parameter \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\theta\:$$\end{document} and the resulting order n of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\sigma\:\tau\:$$\end{document} is given by specific polynomials over \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{F}_{p\:},$$\end{document} as summarized in Table 2. For instance, to construct a Schreier graph for the triangle group \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\left(\mathrm{2,3},16\right)$$\end{document}, one must select a \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\theta\:$$\end{document} that is a root of the polynomial \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\theta\:}^{4}-8{\theta\:}^{3}+20{\theta\:}^{2}-16\theta\:+2\:\equiv\:0\:\left(mod\:p\right)$$\end{document}. The values of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\theta\:$$\end{document} for higher orders n of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\sigma\:\tau\:$$\end{document} can be found by applying the parametrization scheme29. Table 2Relationship between the value of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\theta\:$$\end{document} and the order of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\sigma\:\tau\:$$\end{document}.Polynomial satisfied by \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\theta\:$$\end{document}Order of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\sigma\:\tau\:$$\end{document} \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\theta\:=4$$\end{document} 0 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\theta\:=0$$\end{document} 2 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\theta\:-1=0$$\end{document} 3 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\theta\:-2=0$$\end{document} 4 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\theta\:}^{2}-3\theta\:+1=0$$\end{document} 5 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\theta\:-3=0$$\end{document} 6 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\theta\:}^{3}-5{\theta\:}^{2}+6\theta\:-1=0$$\end{document} 7 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\theta\:}^{2}-4\theta\:+2=0$$\end{document} 8 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\theta\:}^{3}-6{\theta\:}^{2}+9\theta\:-1=0$$\end{document} 9 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\theta\:}^{2}-5\theta\:+5=0$$\end{document} 10 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\theta\:}^{5}-9{\theta\:}^{4}+28{\theta\:}^{3}-35{\theta\:}^{2}+15\theta\:-1=0$$\end{document} 11 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\theta\:}^{2}-4\theta\:+1=0$$\end{document} 12 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\theta\:}^{6}-11{\theta\:}^{5}+45{\theta\:}^{4}-84{\theta\:}^{3}+70{\theta\:}^{2}-21\theta\:+1=0$$\end{document} 13 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\theta\:}^{3}-7{\theta\:}^{2}+14\theta\:-7=0$$\end{document} 14 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\theta\:}^{4}-9{\theta\:}^{3}+26{\theta\:}^{2}-24\theta\:+1=0$$\end{document} 15 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\theta\:}^{4}-8{\theta\:}^{3}+20{\theta\:}^{2}-16\theta\:+2=0$$\end{document} 16 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\theta\:}^{8}-15{\theta\:}^{7}+91{\theta\:}^{6}-286{\theta\:}^{5}+495{\theta\:}^{4}-462{\theta\:}^{3}+210{\theta\:}^{2}-36\theta\:+1=0$$\end{document} 17 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\theta\:}^{4}-9{\theta\:}^{3}+27{\theta\:}^{2}-30\theta\:+9=0$$\end{document} 18 Table 2Relationship between the value of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\theta\:$$\end{document} and the order of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\sigma\:\tau\:$$\end{document}.Polynomial satisfied by \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\theta\:$$\end{document}Order of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\sigma\:\tau\:$$\end{document} \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\theta\:=4$$\end{document} 0 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\theta\:=0$$\end{document} 2 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\theta\:-1=0$$\end{document} 3 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\theta\:-2=0$$\end{document} 4 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\theta\:}^{2}-3\theta\:+1=0$$\end{document} 5 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\theta\:-3=0$$\end{document} 6 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\theta\:}^{3}-5{\theta\:}^{2}+6\theta\:-1=0$$\end{document} 7 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\theta\:}^{2}-4\theta\:+2=0$$\end{document} 8 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\theta\:}^{3}-6{\theta\:}^{2}+9\theta\:-1=0$$\end{document} 9 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\theta\:}^{2}-5\theta\:+5=0$$\end{document} 10 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\theta\:}^{5}-9{\theta\:}^{4}+28{\theta\:}^{3}-35{\theta\:}^{2}+15\theta\:-1=0$$\end{document} 11 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\theta\:}^{2}-4\theta\:+1=0$$\end{document} 12 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\theta\:}^{6}-11{\theta\:}^{5}+45{\theta\:}^{4}-84{\theta\:}^{3}+70{\theta\:}^{2}-21\theta\:+1=0$$\end{document} 13 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\theta\:}^{3}-7{\theta\:}^{2}+14\theta\:-7=0$$\end{document} 14 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\theta\:}^{4}-9{\theta\:}^{3}+26{\theta\:}^{2}-24\theta\:+1=0$$\end{document} 15 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\theta\:}^{4}-8{\theta\:}^{3}+20{\theta\:}^{2}-16\theta\:+2=0$$\end{document} 16 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\theta\:}^{8}-15{\theta\:}^{7}+91{\theta\:}^{6}-286{\theta\:}^{5}+495{\theta\:}^{4}-462{\theta\:}^{3}+210{\theta\:}^{2}-36\theta\:+1=0$$\end{document} 17 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\theta\:}^{4}-9{\theta\:}^{3}+27{\theta\:}^{2}-30\theta\:+9=0$$\end{document} 18 Relationship between the value of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\theta\:$$\end{document} and the order of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\sigma\:\tau\:$$\end{document}. S-box construction The suggested S-box design methodology is founded on Schreier graphs and permutation group theory. This section outlines the process employed to accomplish this task. The Schreier graph adopted in the suggested framework In our S-box construction process, we use the Schreier graph derived from the triangle group \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\left( {2,3,16} \right) = < \sigma \:,\tau \:\:|\:\sigma \:^{2} = \tau \:^{3} = \left( {\sigma \:\tau \:} \right)^{{16}} = 1 >$$\end{document}. This graph is generated by the action of the modular transformation group M on the projective line \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{F}_{257}\cup\:\left\{\infty\:\right\}.$$\end{document} An \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:8\times\:8\:$$\end{document}S-box has 256 unique entries, and 257 is the nearest prime to 256, making \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{F}_{257}$$\end{document} a natural choice for the construction. According to Table 2, for the group \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\left(\mathrm{2,3},16\right),$$\end{document} the parameter \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\theta\:$$\end{document} must satisfy the polynomial\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\:\:{\theta\:}^{4}-8{\theta\:}^{3}+20{\theta\:}^{2}-16\theta\:+2\equiv\:0\:mod\left(257\right)$$\end{document}. We found that \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\theta\:=26$$\end{document} is a root of this polynomial over \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{F}_{257}.$$\end{document}This value serves as the basis for building the corresponding Schreier graph. Setting \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\theta\:=26,$$\end{document} we solve the system of equations \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\left(2.1\right)-\left(2.5\right)$$\end{document} to determine the parameters a,c,d,f,k,r,s and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\varDelta\:,$$\end{document} which define the fractional linear transformations \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\sigma\:$$\end{document} and\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\:\tau\:$$\end{document}. Using \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\varDelta\:=113$$\end{document} in equation \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\left(2.1\right)$$\end{document} gives\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\:r=25$$\end{document}. Substituting k=157 and r=25 into equation \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\left(2.2\right)$$\end{document} yields\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\:s=120$$\end{document}. Then, from equation \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\left(2.3\right)$$\end{document} with d=94 and k=157, we obtain f=167. Finally, putting \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:r=25,\:\:k=157,\:\:s=120,\:\:d=94$$\end{document}, and f=167 in equations \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\left(2.4\right)$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\left(2.5\right)$$\end{document} produces a=81 and c=65. Thus, the final fractional linear mappings are \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\:\sigma\:\left(z\right)=\frac{81z+182}{65z+176}$$\end{document} and\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\:\tau\:\left(z\right)=\frac{94z+5}{167z+162}$$\end{document}. The functions \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\sigma\:$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\tau\:$$\end{document} are applied to every element of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{F}_{257}\cup\:\left\{\infty\:\right\}$$\end{document} to generate their respective permutation cycles. Since these are fractional linear transformations, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\sigma\:\left(z\right)$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\tau\:\left(z\right)$$\end{document} can initially yield fractional values in\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\:{F}_{257}$$\end{document}, so we normalize them by multiplying the numerator and denominator by the modular inverse of the denominator \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\left(\mathrm{m}\mathrm{o}\mathrm{d}\:257\right)$$\end{document}. For instance:\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\sigma\:\left(0\right)=\frac{182}{176}=\left(182\right)\left({176}^{-1}\right)=39\:\:\:\:\left(\mathrm{m}\mathrm{o}\mathrm{d}\:257\right),$$\end{document} and\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\sigma\:\left(1\right)=\frac{6}{241}=\left(6\right)\left({241}^{-1}\right)=96\:\:\:\:\:\:\left(\mathrm{m}\mathrm{o}\mathrm{d}\:257\right).$$\end{document} Similarly, for each\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\:z\in\:{F}_{257}$$\end{document}, the corresponding values of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\sigma\:\left(z\right)$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\tau\:\left(z\right)$$\end{document} can be determined. To keep the main text concise, we present only a representative part of the permutation cycles here. The complete cycle decompositions of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\sigma\:$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\tau\:$$\end{document} are provided in Appendix. For example, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\sigma\:\left(z\right)=\:$$\end{document}(0,39)(1,96)(2,28)(3,15)(4,66)…. and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\tau\:\left(z\right)=\:$$\end{document}(0,27,205)(1,162,195)(2,232,49)(3,14,139)(4,111,242)… . From the above permutation representations, we observe that under the action of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\sigma\:,$$\end{document} the element 0 is mapped to 39 and under\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\:\tau\:,$$\end{document} the image 39 maps further to\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\:63$$\end{document}. Hence, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\:\sigma\:\tau\:\left(0\right)=63$$\end{document}. By repeatedly applying \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\sigma\:$$\end{document}and\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\:\tau\:$$\end{document}, it is found that\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\:{\left(\sigma\:\tau\:\right)}^{16}\left(0\right)=0$$\end{document}. This shows that \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\left(\sigma\:\tau\:\right)}^{16}$$\end{document} fixes the element 0. In fact, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\left(\sigma\:\tau\:\right)}^{16}$$\end{document} fixes all the elements of\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\:{F}_{257}\cup\:\left\{\infty\:\right\}$$\end{document}. This implies that the corresponding Schreier graph satisfies the group relations\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\:{\sigma\:}^{2}={\tau\:}^{3}={\left(\sigma\:\tau\:\right)}^{16}=1$$\end{document}, which define the triangle group \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\left(\mathrm{2,3},16\right).\:$$\end{document}. \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\sigma\:\tau\:=$$\end{document}(0,63,226,168,135,48,84,66,111,202,123,56,70,124,60,55)(1,97,40,137,74,144,7,196,160,222,5,6,235,65,210,43)(2,147,234,206,207,247,52,16,205,68,138,75,172,115,211,169)(3,208,221,28,232,23,236,140,186,126,12,238,61,100,181,166)(4,209,46,31,112,151,231,200,86,26,72,233,189,237,184,248)(8,18,227,190,62,80,183,153,17,223,91,167,83,245,9,133)(10,101,146,128,164,77,44,243,149,212,157,152,88,142,156,89)(11,96,162,143,53,141,220,249,71,159,69,50,116,201,178,34)(13,119,41,102,256,127,244,106,225,85,213,110,171,93,199,\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\infty\:$$\end{document})(14,39,27,32,188,240,90,109,95,252,155,134,87,33,107,15)(19,54,250,216,108,176,21,92,118,47,163,218,98,25,82,254)(20,37,30,81,228,113,51,174,154,177,255,230,148,67,139,136)(22,242,194,204,79,203,224,129,45,121,246,195,59,29,132,150)(24,180,185,173,198,197,105,179,125,78,57,217,117,103,122,229)(35,58,38,161,99,241,131,182,175,192,76,73,145,64,239,214)(36,104,253,219,158,193,215,130,187,114,251,49,165,94,120,191)(42)(170). A part of this Schreier graph is illustrated in Fig. 2, and it consists of 18 orbits under the action of\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\:\sigma\:\tau\:$$\end{document}, as displayed in Figure\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\:3$$\end{document}. Fig. 2Small patch of the coset diagram used in the proposed framework. Prepared using microsoft power point 2021. URL: https://www.microsoft.com/en-us/microsoft-365/powerpoint. Small patch of the coset diagram used in the proposed framework. Prepared using microsoft power point 2021. URL: https://www.microsoft.com/en-us/microsoft-365/powerpoint. Fig. 3Orbits in the coset diagram for the proposed S-box construction. Prepared using microsoft power point 2021. URL: https://www.microsoft.com/en-us/microsoft-365/powerpoint. Orbits in the coset diagram for the proposed S-box construction. Prepared using microsoft power point 2021. URL: https://www.microsoft.com/en-us/microsoft-365/powerpoint. Proposed S-box method The S-box construction is based on the action of the triangle group \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\: < \sigma \:,\tau \:\:|\:\sigma ^{2} = \tau \:^{3} = \left( {\sigma \:\tau \:} \right)^{{16}} = 1 > \:$$\end{document}on the projective line over \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{F}_{257}$$\end{document}. Under this action, the permutation \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\upsigma\:}\tau\:$$\end{document} decomposes into sixteen disjoint cycles of length 16 and two fixed points, collectively covering all 258 elements. Step A: The initial \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:8\times\:8$$\end{document} S-box is constructed from the cycle decomposition of the permutation\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\:{\upsigma\:}\tau\:$$\end{document}. Although the group action is defined on\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\:{F}_{257}\cup\:\left\{\infty\:\right\}$$\end{document}, the elements \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:256\:$$\end{document}and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\infty\:$$\end{document} are excluded, as their presence is only to make the group action possible. The permutation \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\upsigma\:}\tau\:$$\end{document} decomposes into 18 disjoint cycles\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\:\{{\gamma\:}_{1},{\gamma\:}_{2},{\gamma\:}_{3},\dots\:,{\gamma\:}_{18}\}$$\end{document}, of which 16 have length 16 and two are fixed points. Their combined length is\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\:258$$\end{document}, exceeding the required 256 entries of an \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:8\times\:8$$\end{document} S-box by two, which corresponds to the removal of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:256\:$$\end{document}and\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\:\infty\:$$\end{document}. Since the cycles are disjoint, they can be concatenated in arbitrary order, yielding 18! distinct sequences, which reduce to 256-element sequences after removing \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:256\:$$\end{document}and\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\:\infty\:$$\end{document}. Here, 18! represents the theoretical number of possible cycle orderings. In practice, only a selected subset of these orderings was evaluated, and the final arrangement was chosen from the tested candidates on the basis of cryptographic performance. Each sequence is written row-wise into a \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:16\times\:16$$\end{document} matrix to form an S-box. Among all tested arrangements, we adopt the ordering\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\left[{\gamma\:}_{3},{\gamma\:}_{14},{\gamma\:}_{18},{\gamma\:}_{13},{\gamma\:}_{5},{\gamma\:}_{17},{\gamma\:}_{16},{\gamma\:}_{4},{\gamma\:}_{9},{\gamma\:}_{15},{\gamma\:}_{11},{\gamma\:}_{2},{\gamma\:}_{1},{\gamma\:}_{7},{\gamma\:}_{8},{\gamma\:}_{10},{\gamma\:}_{12},{\gamma\:}_{6}\right],$$\end{document} which produces the strongest cryptographic performance. The resulting S-box, shown in Table 3, achieves a nonlinearity of 108. Table 3Initial S-box after step A.Row/Col012345678910111213141502147234206207247521620568138751721152111691241801851731981971051791257857217117103122229217022242194204792032241294512124619559291323150420946311121512312008626722331892371844248423610425321915819321513018711425149165945120191320822128232232361401861261223861100618116613119411021272441062258521311017193199735583816199241131182175192767314564239214819542502161081762192118471632189825822549197401377414471961602225623565210431006322616813548846611120212356701246055111010114612816477442431492121571528814215689121196162143531412202497115969501162011783413143927321882409010995252155134873310715142037308122811351174154177255230148671391361581822719062801831531722391167832459133 Table 3Initial S-box after step A.Row/Col012345678910111213141502147234206207247521620568138751721152111691241801851731981971051791257857217117103122229217022242194204792032241294512124619559291323150420946311121512312008626722331892371844248423610425321915819321513018711425149165945120191320822128232232361401861261223861100618116613119411021272441062258521311017193199735583816199241131182175192767314564239214819542502161081762192118471632189825822549197401377414471961602225623565210431006322616813548846611120212356701246055111010114612816477442431492121571528814215689121196162143531412202497115969501162011783413143927321882409010995252155134873310715142037308122811351174154177255230148671391361581822719062801831531722391167832459133 Initial S-box after step A. Step B: In the second stage, the initial S-box is further enhanced by applying a structured family of permutations generated by the group\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:G={C}_{3}\times\:{C}_{3}\times\:{C}_{15}\times\:{S}_{3}$$\end{document} This group is generated by three permutations a,b, and c acting on the index set \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\{0,\dots\:,255\}.$$\end{document} Their cycle decompositions are:\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:a=\left(\mathrm{5,251,106}\right)\left(\mathrm{31,246,114}\right)\left(\mathrm{36,143,60}\right),$$\end{document}\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:b=\left(\mathrm{5,251}\right)\left(\mathrm{53,83,162}\right)\left(\mathrm{8,242,78}\right)\left(\mathrm{12,26,109}\right)\left(\mathrm{18,112,146}\right)\left(\mathrm{42,147,227}\right),$$\end{document}\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:c=\left(\mathrm{5,251}\right)\left(\mathrm{64,76,238,111,67}\right)\left(\mathrm{71,199,237}\right)\left(\mathrm{98,115,256}\right).$$\end{document} GAP is used here as a computational algebra system to verify the defining relations of the generated permutation group and to confirm its algebraic structure. According to GAP, the group G admits the presentation:\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\: < a,b,c:a^{3} = b^{6} = c^{{30}} = bcb^{{ - 1}} c^{{ - 1}} = ab^{2} a^{{ - 1}} b^{{ - 2}} = acba^{{ - 1}} c^{{ - 1}} b^{{ - 1}} = ab^{{ - 1}} aca^{{ - 1}} c^{{ - 1}} a^{{ - 1}} b = 1 > ,$$\end{document} and is a non-abelian group of order 810 Every element of G only acts on the positions of S-box entries, but not the values; thus, bijectivity is always maintained. Suppose that \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{S}_{0}$$\end{document} is the 256-element first S-box generated by the triangle-group cycles. We flatten the S-box into a vector:\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{S}_{0}=\left\{{s}_{0},{s}_{1},{s}_{2},\dots\:,{s}_{255}\right\}.$$\end{document} The action of any permutation \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:g\in\:G$$\end{document} is:\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{S}_{g}\left[i\right]={S}_{0}\left[{g}^{-1}\left(i\right)\right],$$\end{document} i.e., the value that was initially in position \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{g}^{-1}\left(i\right)$$\end{document} of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{S}_{0}$$\end{document} is moved to position i of the new S-box. This generates a systematic positional rearrangement that preserves all 256 distinct output values. G is non-abelian, so its 810 elements generate a large set of rearrangements, allowing an extensive search space of possible S-box candidates. All group elements were applied to \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{S}_{0}$$\end{document}, and each of the resulting S-boxes was tested with nonlinearity, SAC, BIC, differential uniformity, and linear approximation probability. Of all candidates, the permutation \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{a}^{2}{b}^{5}{c}^{17}$$\end{document} gives the best overall cryptographic performance, with an average nonlinearity of 111.25. This permutation results in an S-box that is chosen as the proposed S-box and is shown in Table 4. Table 4Proposed S-box after step B.Row/Col012345678910111213141502147234206225247524920568138110172115211169124971851731981971051791257557217117103582292170222427220479203224129301212461955929132315042094663112151231200862682233189237364248429310425321914819321513018718425118165945120191312082212823223236140186126122386110061811331311941102127244106918521378171671807358016616199241131182175192767314564239214819542502161081762192118471632189825194254911994513774144719616022256235652104310032261681354884661112021235670124605511101011461281647744243149212157152881421568912119616214353141158249711596950116201178341314392732188240901099525215513487331071514203740812281135117415417725523022011413913615816227190621221831531722320716783245938 Table 4Proposed S-box after step B.Row/Col012345678910111213141502147234206225247524920568138110172115211169124971851731981971051791257557217117103582292170222427220479203224129301212461955929132315042094663112151231200862682233189237364248429310425321914819321513018718425118165945120191312082212823223236140186126122386110061811331311941102127244106918521378171671807358016616199241131182175192767314564239214819542502161081762192118471632189825194254911994513774144719616022256235652104310032261681354884661112021235670124605511101011461281647744243149212157152881421568912119616214353141158249711596950116201178341314392732188240901099525215513487331071514203740812281135117415417725523022011413913615816227190621221831531722320716783245938 Proposed S-box after step B. For the convenience of the reader, the workflow of the proposed S-box construction is illustrated in Fig. 4. For reproducibility, Figs. 1, 2, 3 and 4 were prepared using Microsoft PowerPoint 2021, Microsoft Corporation, Redmond, WA, USA, URL: https://www.microsoft.com/en-us/microsoft-365/powerpoint, while Fig. 5 was generated using MATLAB R2024a, The MathWorks, Inc., Natick, MA, USA, URL: https://www.mathworks.com/products/matlab.html. Algebraic analysis A detailed performance evaluation of the proposed S-box is conducted using modern cryptographic metrics. The scheme is shown to be highly resistant to algebraic manipulations and to have exceptional structural properties, which are confirmed by the analysis. Nonlinearity Nonlinearity is one of the basic requirements for assessing the resistance of an S-box to linear and algebraic cryptanalysis. The nonlinearity \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{N}_{f}$$\end{document} of a Boolean function \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:f:{\left\{\mathrm{0,1}\right\}}^{n}\to\:\left\{\mathrm{0,1}\right\}$$\end{document} defines the minimum Hamming distance between f and the set of all affine functions, that is,\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{N}_{f}=\underset{g\in\:A}{\mathrm{min}}{d}_{H}\left(f,g\right),$$\end{document} where A refers to the set of affine Boolean functions30. The nonlinearity results of the proposed S-box are reported in Table 5. The component Boolean functions attain values of 110 and 112, yielding an average nonlinearity of 111.25 and a minimum value of 110, which confirms strong resistance to linear cryptanalysis. Table 5The details of the nonlinearities of Boolean functions in the proposed S-box. \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{f}_{1}$$\end{document} \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{f}_{2}$$\end{document} \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{f}_{3}$$\end{document} \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{f}_{4}$$\end{document} \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{f}_{5}$$\end{document} \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{f}_{6}$$\end{document} \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{f}_{7}$$\end{document} \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{f}_{8}$$\end{document} 112 110 112 112 112 112 110 110 Table 5The details of the nonlinearities of Boolean functions in the proposed S-box. \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{f}_{1}$$\end{document} \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{f}_{2}$$\end{document} \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{f}_{3}$$\end{document} \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{f}_{4}$$\end{document} \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{f}_{5}$$\end{document} \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{f}_{6}$$\end{document} \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{f}_{7}$$\end{document} \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{f}_{8}$$\end{document} 112 110 112 112 112 112 110 110 The details of the nonlinearities of Boolean functions in the proposed S-box. Fig. 4Workflow of the proposed S-box construction scheme. Prepared using microsoft power point 2021. URL: https://www.microsoft.com/en-us/microsoft-365/powerpoint. Workflow of the proposed S-box construction scheme. Prepared using microsoft power point 2021. URL: https://www.microsoft.com/en-us/microsoft-365/powerpoint. Differential uniformity Differential Uniformity \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\left(DU\right)$$\end{document} is a primary metric for evaluating an S-box’s resistance to differential cryptanalysis. For any non-zero input difference \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\varDelta\:x\in\:{\left\{\mathrm{0,1}\right\}}^{n}$$\end{document} and output difference \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\varDelta\:y\in\:{\left\{\mathrm{0,1}\right\}}^{m},$$\end{document} the differential uniformity of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:f\:$$\end{document}is defined as:\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\delta \:\left( f \right) = \mathop {\max }\limits_{{\Delta x \ne \:0,\Delta \:y}} \left( {\# \left\{ {x \in \:\left\{ {0,1} \right\}^{n} |f\left( x \right) \otimes f\left( {x \otimes \Delta \:x} \right) = \Delta \:y} \right\}} \right).$$\end{document} A smaller value of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\delta\:\left(f\right)$$\end{document} indicates stronger resistance, as it bounds the maximum number of occurrences of any input–output difference pair. Table 6 illustrates the differential uniformity results of the proposed S-box. Table 6Distribution of values in the differential distribution table for the proposed S-box.10.06.06.06.06.06.08.08.08.08.06.010.06.08.08.06.08.010.08.06.06.08.06.08.08.06.06.08.08.08.06.06.06.06.06.08.08.06.06.06.06.06.06.06.06.06.06.08.06.08.08.06.06.06.08.010.06.08.08.06.08.06.06.06.06.06.06.06.06.06.06.06.08.06.08.06.06.06.08.08.06.010.06.06.04.010.06.08.08.06.06.06.06.08.08.06.08.06.06.08.06.010.06.06.06.06.06.06.06.06.08.06.06.08.06.06.010.06.06.08.06.06.06.06.08.012.06.06.08.08.06.08.08.06.06.06.08.08.06.06.06.06.06.06.06.06.06.06.08.08.06.08.08.06.08.06.06.08.06.08.06.06.06.06.06.06.06.06.06.06.06.06.06.06.08.06.06.08.06.08.06.06.06.06.04.06.08.08.06.06.06.06.08.06.06.08.06.06.06.06.06.06.06.08.06.06.08.06.06.08.08.06.06.06.06.08.06.08.06.08.010.06.06.06.06.06.08.06.08.08.06.06.06.08.06.08.06.06.06.06.06.08.06.06.08.06.08.010.06.06.08.06.06.06.08.00.0 Table 6Distribution of values in the differential distribution table for the proposed S-box.10.06.06.06.06.06.08.08.08.08.06.010.06.08.08.06.08.010.08.06.06.08.06.08.08.06.06.08.08.08.06.06.06.06.06.08.08.06.06.06.06.06.06.06.06.06.06.08.06.08.08.06.06.06.08.010.06.08.08.06.08.06.06.06.06.06.06.06.06.06.06.06.08.06.08.06.06.06.08.08.06.010.06.06.04.010.06.08.08.06.06.06.06.08.08.06.08.06.06.08.06.010.06.06.06.06.06.06.06.06.08.06.06.08.06.06.010.06.06.08.06.06.06.06.08.012.06.06.08.08.06.08.08.06.06.06.08.08.06.06.06.06.06.06.06.06.06.06.08.08.06.08.08.06.08.06.06.08.06.08.06.06.06.06.06.06.06.06.06.06.06.06.06.06.08.06.06.08.06.08.06.06.06.06.04.06.08.08.06.06.06.06.08.06.06.08.06.06.06.06.06.06.06.08.06.06.08.06.06.08.08.06.06.06.06.08.06.08.06.08.010.06.06.06.06.06.08.06.08.08.06.06.06.08.06.08.06.06.06.06.06.08.06.06.08.06.08.010.06.06.08.06.06.06.08.00.0 Distribution of values in the differential distribution table for the proposed S-box. Strict avalanche criterion The Strict Avalanche Criterion (SAC) measures the sensitivity of a Boolean function to single-bit input changes. A Boolean function f satisfies SAC if, for every vector y with Hamming weight\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\:{w}_{H}\left(y\right)=1$$\end{document}, the function \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:f\left(x\right)⨁f\left(x⨁y\right)\:$$\end{document}is balanced, meaning that each output bit changes with probability\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\:\raisebox{1ex}{$1$}\!\left/\:\!\raisebox{-1ex}{$2$}\right.$$\end{document}. The SAC results for the proposed S-box are presented in Table 7. The ideal value for each entry is\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\:0.5$$\end{document}, and all observed values lie close to the ideal value of\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\:0.5$$\end{document}, demonstrating strong avalanche behavior and effective diffusion. Table 7Strict avalanche criterion (SAC) matrix for the proposed S-box.0.59380.45310.50.51560.43750.53120.51560.50.51560.56250.54690.46880.53120.48440.50.53120.48440.56250.54690.48440.46880.53120.53120.51560.50.54690.39060.50.45310.59380.43750.50.51560.46880.43750.46880.51560.51560.59380.51560.48440.50.54690.48440.48440.53120.53120.50.50.59380.51560.51560.45310.48440.45310.46880.43750.53120.43750.57810.51560.57810.48440.5 Table 7Strict avalanche criterion (SAC) matrix for the proposed S-box.0.59380.45310.50.51560.43750.53120.51560.50.51560.56250.54690.46880.53120.48440.50.53120.48440.56250.54690.48440.46880.53120.53120.51560.50.54690.39060.50.45310.59380.43750.50.51560.46880.43750.46880.51560.51560.59380.51560.48440.50.54690.48440.48440.53120.53120.50.50.59380.51560.51560.45310.48440.45310.46880.43750.53120.43750.57810.51560.57810.48440.5 Strict avalanche criterion (SAC) matrix for the proposed S-box. Bit independence criterion (BIC) Webster and Tavares31 proposed the Bit Independence Criterion (BIC), which is a necessary condition in the evaluation of the security of cryptographic S-boxes. It provides that the avalanche variables of the various output bits must be statistically independent. This property is assessed by calculating the correlation coefficients of all pairs of avalanche variables, which quantify their level of dependence. Formally, a vectorial Boolean function \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:F=\left({f}_{1},{f}_{2},\dots\:,{f}_{n}\right):{\left\{\mathrm{0,1}\right\}}^{n}\to\:{\left\{\mathrm{0,1}\right\}}^{n}$$\end{document} satisfies the BIC: given two different output bits \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{f}_{j}$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{f}_{k}\:\left(j\ne\:k\right)$$\end{document}, the Boolean function \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{f}_{j}⨁{f}_{k}$$\end{document} satisfies both the strict avalanche criterion and high nonlinearity. This guarantees that variations in one input bit cause independent and unpredictable variations in other output bits, which increases diffusion. Tables 8 and 9 give the BIC results of the proposed S-box in terms of SAC and nonlinearity, respectively. The behavior observed demonstrates that the output bits are statistically independent and highly nonlinear when considered together, which meets the bit independence criterion. Table 8BIC outcomes for SAC related to proposed S-box.00.50390.50980.50.51170.50590.51370.52930.503900.51370.46290.48830.47660.48240.51950.50980.513700.50.51560.52540.49220.50780.50.46290.500.51760.50.51560.5020.51170.48830.51560.517600.50980.48830.51560.50590.47660.52540.50.509800.49410.51370.51370.48240.49220.51560.48830.494100.49020.52930.51950.50780.5020.51560.51370.49020 Table 8BIC outcomes for SAC related to proposed S-box.00.50390.50980.50.51170.50590.51370.52930.503900.51370.46290.48830.47660.48240.51950.50980.513700.50.51560.52540.49220.50780.50.46290.500.51760.50.51560.5020.51170.48830.51560.517600.50980.48830.51560.50590.47660.52540.50.509800.49410.51370.51370.48240.49220.51560.48830.494100.49020.52930.51950.50780.5020.51560.51370.49020 BIC outcomes for SAC related to proposed S-box. Table 9BIC outcomes for nonlinearity related to proposed S-box.01041041041041041001041040961041021049810410496010210410210210610410410209610410610410410210496010810410610410410210410801021041009810210610410201021041041061041061041020 Table 9BIC outcomes for nonlinearity related to proposed S-box.01041041041041041001041040961041021049810410496010210410210210610410410209610410610410410210496010810410610410410210410801021041009810210610410201021041041061041061041020 BIC outcomes for nonlinearity related to proposed S-box. Linear approximation probability (LAP) Another aim of S-box design is to create a highly nonlinear connection between the input and output of a cipher, which produces ciphertext that appears random to an attacker. Linear cryptanalysis can be used to exploit any linearity in this mapping. The Linear Approximation Probability (LAP)32 is defined as follows:\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:LAP=\underset{\alpha\:,\beta\:\ne\:0}{\mathrm{max}}\left|\frac{\#\left\{x\in\:X|x.\alpha\:=S\left(x\right).\beta\:\right\}}{{2}^{n}}-\frac{1}{2}\right|,$$\end{document} where \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\alpha\:$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\beta\:\:$$\end{document}are the input and output masks, respectively. A smaller LAP means that the S-box is more resistant to linear attacks. In the case of the proposed S-box, the theoretical maximum LAP is 0.5, but the actual maximum is 0.1328, which is significantly lower than 0.5, indicating that there is very little bias and a high resistance to linear cryptanalysis. Comparison analysis Through the use of group actions and a new refinement process, we have created an S-box with an average nonlinearity score of 111.25, exceeding most modern designs. Table 10 confirms through comparative analysis that although other S-boxes might be good in one particular area, our design provides a more balanced cryptographic profile, with outstanding resistance to a broad spectrum of attacks. Table 10Comparison of nonlinearity, bit independence, avalanche criteria, differential uniformity, and linear probability of the proposed S-box with existing methods.S-boxAverage NLBIC-SACSACBIC-NLDULPProposed S-box111.250.50380.5059103.00120.1328Initial S-box1080.49450.5022103.29120.1406Ahmed9107.500.49820.4944104.35100.1250Hayat12106.250.49750.5112103.93120.1484Ibrahim33106.50.49870.5010103.93100.125Alshammar341070.50190.4949102.29120.141Alhadaw35108.50.50110.4995103.85100.109Long36109.750.49870.5042110.660.0859Soto37106.50.50190.4943103.35120.1468Yan38105.50.50310.5065103.57100.1328Zhou391070.50500.4993103.29100.1328Jama401080.50200.5029104.43140.1328Wang411050.50020.5063104.07100.1328Ejaz421070.49850.5078104.43120.1406Zaml43109.750.50410.4998104.14100.1171Murtaza44106.750.49970.5051104.43100.1406Farah45106.500.49830.4995104.57100.1171Alhadawi46108.50.50480.491103.78100.0791Özkaynak47103.200.50090.5048103.70100.1289Aydin48106.750.50430.4995105.07120.1289 Table 10Comparison of nonlinearity, bit independence, avalanche criteria, differential uniformity, and linear probability of the proposed S-box with existing methods.S-boxAverage NLBIC-SACSACBIC-NLDULPProposed S-box111.250.50380.5059103.00120.1328Initial S-box1080.49450.5022103.29120.1406Ahmed9107.500.49820.4944104.35100.1250Hayat12106.250.49750.5112103.93120.1484Ibrahim33106.50.49870.5010103.93100.125Alshammar341070.50190.4949102.29120.141Alhadaw35108.50.50110.4995103.85100.109Long36109.750.49870.5042110.660.0859Soto37106.50.50190.4943103.35120.1468Yan38105.50.50310.5065103.57100.1328Zhou391070.50500.4993103.29100.1328Jama401080.50200.5029104.43140.1328Wang411050.50020.5063104.07100.1328Ejaz421070.49850.5078104.43120.1406Zaml43109.750.50410.4998104.14100.1171Murtaza44106.750.49970.5051104.43100.1406Farah45106.500.49830.4995104.57100.1171Alhadawi46108.50.50480.491103.78100.0791Özkaynak47103.200.50090.5048103.70100.1289Aydin48106.750.50430.4995105.07120.1289 Comparison of nonlinearity, bit independence, avalanche criteria, differential uniformity, and linear probability of the proposed S-box with existing methods. As shown in Table 10, the proposed S-box has a balanced set of cryptographic strengths compared to recent designs. It has one of the highest nonlinearities in the Table 10 at 111.25 and exceeds most of its competitors, such as Ahmed12, Hayat9, and Ibrahim33. This degree of nonlinearity is only approached by a few methods, including Long36 and Zamli43. The differential uniformity of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:12\:$$\end{document}is comparable to the performance of a number of known S-boxes and is within the secure range of current block-cipher design. The linear approximation probability of 0.1328 is competitive and better than the original S-box, indicating that resistance to linear cryptanalysis is improved after the refinement step performed with the help of the permutation group \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{C}_{3}\times\:{C}_{3}\times\:{C}_{15}\times\:{S}_{3}$$\end{document}. The SAC and BIC values are near the optimal values, which demonstrate that single-bit variations in the input spread uniformly through the output and that the output bits act independently. Recent algebraic image-encryption schemes based on non-chain rings49, DNA encoding50, and Henon map51 typically combine S-box construction with additional permutation, diffusion, or XOR layers. In such frameworks, the reported security behaviour reflects the combined effect of the full encryption pipeline rather than the isolated strength of a single 8 × 8 S-box. By contrast, the present work focuses on the nonlinear component itself and shows that the proposed S-box alone already attains an average nonlinearity of 111.25. The main advantage of the present framework is therefore its mathematical transparency and reproducibility: the construction follows from an explicit group action, the intermediate cycle structure is interpretable, and the refinement stage is carried out in a well-defined non-abelian search space. We do not claim that this makes the method universally superior to all algebraic image ciphers under every system-level metric; rather, it shows that strong and balanced cryptographic behaviour can already be achieved at the level of the S-box itself, without relying on auxiliary chaotic or DNA-based stages. At the same time, the dynamic and key-dependent features used in some algebraic image ciphers remain an interesting direction for future extensions of the present framework. Texture-based analysis of s-box substitution This image-based assessment is further supported by recent studies on secure visual data processing, where image encryption, privacy protection, and watermarking, are commonly used to evaluate the reliability of modern image-security frameworks52–54. To evaluate an S-box’s effect on image data, researchers often use the majority logic criterion (MLC), which looks at standard texture features. These include contrast, correlation, energy, homogeneity, and entropy. Each captures a different aspect of an image’s texture: Contrast measures local intensity change; large contrast implies large intensity differences between adjacent pixels.Correlation is used to represent the frequency of encountering particular pixel pairs (i.e., the dependency of neighboring pixels).Energy (or angular second moment) is the sum of squared GLCM values, which shows uniformity; higher energy indicates a more uniform repeating pattern.Homogeneity (inverse difference moment) determines the proximity of pixel pairs to the diagonal of the GLCM, which indicates smoothness.Entropy is a measure of the randomness of the grey-level distribution; the greater the entropy, the more random and less predictable the image is. Entropy may be as large as 8 (in the case of an 8-bit image), with values close to 8 indicating a very high level of randomness. Contrast measures local intensity change; large contrast implies large intensity differences between adjacent pixels. Contrast measures local intensity change; large contrast implies large intensity differences between adjacent pixels. Correlation is used to represent the frequency of encountering particular pixel pairs (i.e., the dependency of neighboring pixels). Correlation is used to represent the frequency of encountering particular pixel pairs (i.e., the dependency of neighboring pixels). Energy (or angular second moment) is the sum of squared GLCM values, which shows uniformity; higher energy indicates a more uniform repeating pattern. Energy (or angular second moment) is the sum of squared GLCM values, which shows uniformity; higher energy indicates a more uniform repeating pattern. Homogeneity (inverse difference moment) determines the proximity of pixel pairs to the diagonal of the GLCM, which indicates smoothness. Homogeneity (inverse difference moment) determines the proximity of pixel pairs to the diagonal of the GLCM, which indicates smoothness. Entropy is a measure of the randomness of the grey-level distribution; the greater the entropy, the more random and less predictable the image is. Entropy may be as large as 8 (in the case of an 8-bit image), with values close to 8 indicating a very high level of randomness. Entropy is a measure of the randomness of the grey-level distribution; the greater the entropy, the more random and less predictable the image is. Entropy may be as large as 8 (in the case of an 8-bit image), with values close to 8 indicating a very high level of randomness. Substitution-only experiment setup We isolate the S-box effect by running a substitution-only experiment. Using standard 8-bit grayscale images (Peppers, Rice, and Fabric), each \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:512\times\:512$$\end{document} pixels, we applied the S-box to every pixel independently (so output \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:C\left(i,j\right)=S\left(P\left(i,j\right)\right)$$\end{document}. No other mixing (permutation or diffusion) is applied. Then we analyze texture via GLCMs. The steps are: i.Convert each test image to 8-bit grayscale, size \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:512\times\:512$$\end{document}.ii.For each S-box under test, replace every pixel value \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:P\left(i,j\right)$$\end{document} by \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:C\left(i,j\right)=S\left(P\left(i,j\right)\right)$$\end{document}.iii.Build GLCMs with 256 gray levels, using a pixel offset of distance 1 in four directions: horizontal \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:0^\circ\:,\:45^\circ\:,$$\end{document} vertical \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:90^\circ\:,$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:135^\circ\:.$$\end{document} These correspond to offsets \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\left[\mathrm{0,1}\right],\:\left[-\mathrm{1,1}\right],\:\left[-\mathrm{1,0}\right],\:[-1,-1]$$\end{document}.iv.For each direction, make the GLCM symmetric by adding it to its transpose. This counts each gray-level pair both ways.v.From each directional GLCM, compute contrast, correlation, energy, and homogeneity. Then average each feature over the four directions to get one value per image and S-box. i.Convert each test image to 8-bit grayscale, size \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:512\times\:512$$\end{document}. Convert each test image to 8-bit grayscale, size \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:512\times\:512$$\end{document}. ii.For each S-box under test, replace every pixel value \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:P\left(i,j\right)$$\end{document} by \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:C\left(i,j\right)=S\left(P\left(i,j\right)\right)$$\end{document}. For each S-box under test, replace every pixel value \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:P\left(i,j\right)$$\end{document} by \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:C\left(i,j\right)=S\left(P\left(i,j\right)\right)$$\end{document}. iii.Build GLCMs with 256 gray levels, using a pixel offset of distance 1 in four directions: horizontal \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:0^\circ\:,\:45^\circ\:,$$\end{document} vertical \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:90^\circ\:,$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:135^\circ\:.$$\end{document} These correspond to offsets \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\left[\mathrm{0,1}\right],\:\left[-\mathrm{1,1}\right],\:\left[-\mathrm{1,0}\right],\:[-1,-1]$$\end{document}. Build GLCMs with 256 gray levels, using a pixel offset of distance 1 in four directions: horizontal \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:0^\circ\:,\:45^\circ\:,$$\end{document} vertical \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:90^\circ\:,$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:135^\circ\:.$$\end{document} These correspond to offsets \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\left[\mathrm{0,1}\right],\:\left[-\mathrm{1,1}\right],\:\left[-\mathrm{1,0}\right],\:[-1,-1]$$\end{document}. iv.For each direction, make the GLCM symmetric by adding it to its transpose. This counts each gray-level pair both ways. For each direction, make the GLCM symmetric by adding it to its transpose. This counts each gray-level pair both ways. v.From each directional GLCM, compute contrast, correlation, energy, and homogeneity. Then average each feature over the four directions to get one value per image and S-box. From each directional GLCM, compute contrast, correlation, energy, and homogeneity. Then average each feature over the four directions to get one value per image and S-box. This mirrors standard practice (for example, MATLAB’s graycoprops does the averaging across GLCMs). Retaining only the substitution step allows us to directly observe the impact of the S-box on texture disruption. The quantitative findings are summarized in Table 11. Figure 5 demonstrates the original images and the substitution-only transformed versions for visual comparison. Table 11Analysis of the encrypted images in terms of their texture features using the proposed S-box.ImageS-boxEntropyContrastCorrelationEnergyHomogeneityPeppersOriginal6.991149.6320.988570.00200950.4617Proposed6.99117849.70.191310.00200950.21484AES6.991110,5670.1830.00200950.21792Hayat126.99118586.30.173890.00200950.21949RiceOriginal6.987781.0540.977260.000663020.34153Proposed6.98778393.20.126310.000663020.13886AES6.987710,0610.105430.000663020.14261Hayat126.987710,0470.0988060.000663020.13923FabricOriginal7.3135363.260.888950.000184170.19654Proposed7.31359310.10.0552130.000184170.074234AES7.313511,3780.0438570.000184170.074713Hayat147.313510,4260.0330060.000184170.07375 Table 11Analysis of the encrypted images in terms of their texture features using the proposed S-box.ImageS-boxEntropyContrastCorrelationEnergyHomogeneityPeppersOriginal6.991149.6320.988570.00200950.4617Proposed6.99117849.70.191310.00200950.21484AES6.991110,5670.1830.00200950.21792Hayat126.99118586.30.173890.00200950.21949RiceOriginal6.987781.0540.977260.000663020.34153Proposed6.98778393.20.126310.000663020.13886AES6.987710,0610.105430.000663020.14261Hayat126.987710,0470.0988060.000663020.13923FabricOriginal7.3135363.260.888950.000184170.19654Proposed7.31359310.10.0552130.000184170.074234AES7.313511,3780.0438570.000184170.074713Hayat147.313510,4260.0330060.000184170.07375 Analysis of the encrypted images in terms of their texture features using the proposed S-box. Table 11 shows that the suggested S-box significantly alters the texture of all test images. The contrast values in the original images are relatively low (e.g., 49.632 in the case of Peppers), and correlation is extremely high (maximum 0.98857), which means that there is a high similarity between adjacent pixels. Following the implementation of the suggested S-box, contrast significantly increases (e.g., 7849.7 in the case of Peppers), and correlation reduces significantly (to 0.19131), which indicates a high level of reduction in pixel dependency. It should be noted that the entropy does not change in all S-boxes. The reason is that the substitution operation is bijective and leaves the histogram of pixel intensities unchanged; thus, entropy, which is only sensitive to the gray-level distribution, is also invariant under substitution-only transformation. The same can be said about the Rice and Fabric images, in which the contrast increases to 8393.2 and 9310.1, respectively, and the correlation reduces to 0.12631 and 0.055213, respectively. Concurrently, the homogeneity values also decrease significantly in all images (e.g., from 0.4617 to 0.21484 in Peppers), which means that smooth texture structures are lost. Such transformations confirm that the given S-box is indeed very effective in breaking local intensity associations and introducing a higher degree of randomness into the image. By comparison, the AES and Hayat S-boxes have similar behavior in all three images. In certain instances, their contrast values are slightly higher, while correlation and homogeneity show similar decreasing trends. Fig. 5Original and substitution-only transformed images. Generated using MATLAB R2024a. URL: https://www.mathworks.com/products/matlab.html. Substitution only: Proposed S-box, Substitution only: AES, Substitution only: Hayat12. Original and substitution-only transformed images. Generated using MATLAB R2024a. URL: https://www.mathworks.com/products/matlab.html. Substitution only: Proposed S-box, Substitution only: AES, Substitution only: Hayat12. Conclusion In this paper, we present a new method for constructing a cryptographically sound \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:8\times\:8$$\end{document} S-box by utilizing both group theory and graph theory. The technique described works by applying the modular transformation group to the projective line over a finite field, and the resulting Schreier graphs, together with their strong structural properties, are then used to derive permutation cycles, which are used to construct S-boxes. To further enhance the cryptographic strength, a refinement process is applied based on a non-abelian permutation group, ensuring improved diffusion and randomness without loss of bijectivity. The proposed S-box has been tested with respect to standard cryptographic indicators, including nonlinearity, differential uniformity, strict avalanche criterion, bit-independence criterion, and linear approximation probability. The results show that the designed S-box is nonlinear and resistant to both linear and differential cryptanalysis, and its performance is balanced and competitive with existing methods. Furthermore, the use of texture-based analysis via substitution-only image experiments demonstrates that the proposed S-box enhances randomness and reduces statistical dependencies in image data. These findings indicate that group-theoretic constructions are a promising direction for S-box design with strong mathematical security guarantees. One limitation is that the current design is neither dynamic nor key-dependent; however, variants of the design can be explored in future work. Another limitation is that the image analysis is based on substitution-only experiments, which are useful for isolating the effect of the S-box but do not represent a complete image-encryption system. Therefore, future work may extend the proposed framework toward key-dependent S-box generation, dynamic variants, and full cipher-level security evaluation.