Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 31 No. 4s (2024) 672 https://internationalpubls.com Geometric Insights into Non-Euclidean Spaces: A Study in Hyperbolic Geometry 1S. Balamuralitharan, 2A. Nasreen Sultana, 3R. Arulprakasam, 4V.Venkata Kumar, 5R Giri Prasad 1Adjunct Faculty, Department of Pure and Applied Mathematics, Saveetha School of Engineering, SIMATS, Chennai, Tamil Nadu, India Email Id: balamurali.maths@gmail.com 2Assistant Professor, Department of Mathematics, S.D.N.B. Vaishnav College for Women, Chennai, Tamil Nadu, India Email : a.nasreensultana@gmail.com 3Department of Mathematics, College of Engineering and Technology, SRM Institute of Science and Technology, SRM Nagar, Kattankulathur - 603203, Chengalpattu District, Tamilnadu, India r.aruljeeva@gmail.com 4Associate Professor, Department of Mathematics, Aditya University, Surampalem, India, venkatakumar.v@aec.edu.in 5Associate Professor, Dept of Petroleum Technology, Aditya University, Surampalem, India, hod_pt@aec.edu.in Article History: Received: 12-02-2024 Revised: 15-03-2024 Accepted: 19-04-2024 Abstract: This work evaluates the core principles along with models and usages of hyperbolic geometry which establishes itself as a main category of non-Euclidean geometry because it lacks Euclid's parallel axiom. The analysis of lines angles and triangles in hyperbolic spaces takes place through the study of the Poincaré disk model and the hyperboloid model. This paper examines the separation between Euclidean structures and hyperbolic structures while showing their distinctive properties which affect mathematical and physical applications. Keywords— Hyperbolic Geometry, Non-Euclidean Space, Poincaré Disk Model, Hyperboloid Model, Parallel Postulate, Curved Spaces, Geometric Structures I. INTRODUCTION Through geometry humans have historically established one of their first methods to understand the world since ancient times. The foundation for understanding geometry as a discipline which developed by Euclid in 300 BCE originated first from useful requirements like engineering and farming land. The system consisted of five postulates where the fifth discussed parallel lines specifically stating that one and only one line exists parallel to any given line. Although fundamental to the structure the postulate presented difficulties which made it seem more complicated and less straightforward than other principles. Mathematicians fruitlessly tried to demonstrate the parallel postulate as a theorem for more than two thousand Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 31 No. 4s (2024) 673 https://internationalpubls.com years while treating it as a basic starting assumption. The unsuccessful attempts at proving the parallel postulate throughout history finally revealed in the 19th century that geometries could exist which rejected this assumption [1-4]. Mathematicians discovered the new era of non-Euclidean geometries because of which hyperbolic geometry became a major research area. From a point outside a given line in hyperbolic space an infinite number of lines will extend without intersecting the first line. Within hyperbolic geometry we find a world that breaks most of the fundamental principles associated with Euclidean space. In hyperbolic geometry one finds the sum of triangle angles to be smaller than 180 degrees and the circumference-to-diameter ratio surpasses π while geodesics display curvilinear shapes based on visualization approaches. Multiple methods exist for displaying hyperbolic geometry features which enhance its difficult concepts for better understanding. Two essential models exist in both Poincaré disk and hyperboloid. When represented by the Poincaré disk model hyperbolic space exists inside a Euclidean circle and its lines exhibit circular shapes that intersect the border circle while maintaining 90-degree angles. The model maintains angle relationships as it acts conformally but results in dramatic distance modifications close to its boundary. The hyperboloid model of hyperbolic space implements a visualization through a double-horned hyperbolic surface which exists inside Minkowski space to present algebraic connections [15]. Hyperbolic geometry finds its use in theoretical subjects as well as practical applications. The physics field uses hyperbolic geometry to explain the large-scale organization of the universe so long as the space has negative curvature. Complex networks together with data science use hyperbolic spaces to create efficient representations of hierarchical structures where social networks or knowledge graphs benefit from better expansion models than Euclidean spaces. The infinite imagery created by M.C. Escher became possible through his representation of hyperbolic principles in artwork. The enhanced relevance of hyperbolic geometry makes it hard to learn and understand at an intuitive level since it violates our natural visual-spatial instincts which developed from living in a flat Euclidean environment. The research community strives to accomplish two essential goals: enhanced theoretical insights and innovative approaches towards visualization, simulation as well as application of hyperbolic spaces [13-14]. This research examines fundamental hyperbolic geometry models while studying essential properties between parallelism and triangle structure and distance computation along with their visual representation through computational tools. We use hyperbolic concepts to contrast with Euclidean concepts to build both conceptual and physical understanding of hyperbolic spaces. Our study includes an investigation of modern scientific and technological areas where hyperbolic geometry proves crucial along with existing and anticipated applications. Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 31 No. 4s (2024) 674 https://internationalpubls.com Novelty and Contribution The presented research presents modern computational methods connected to classical theory for studying hyperbolic geometry. The fundamental concepts of hyperbolic geometry already exist yet researchers have not yet identified methods to link formal theories to visual understandings together with applications across different fields. Our contribution is threefold: • The first component analyzes both major models including Poincaré disk and hyperboloid model through precise mathematical comparison along with visualization and practical computational assessments. Each model finds unique application requirements because they preserve angles differently from maintaining metric linearity depending on the given task. • Our methodology utilizes computer-generated visualizations that enhance representation of hyperbolic properties. The utilization of contemporary software enables the demonstration of hyperbolic lines and geodesics alongside tessellations through visualizations which simplify the complex nature of hyperbolic space especially for those who are not mathematical experts. The methodology creates a connection which allows complex theoretical information to meet functional learning opportunities. • The discussion about hyperbolic geometry includes new applications which cover network science and data visualization. This study demonstrates that hyperbolic geometry serves essential practical purposes because we can use its properties to represent hierarchical datasets in practical applications. Our research seeks dual contributions to studies about hyperbolic geometry while working to establish its widespread use across academic fields and practical applications. We recommend using an integrated strategy to explain hyperbolic spaces while opening new directions for scientific development and research [7]. II. RELATED WORKS In 2020 T. Novello et.al., V. Da Silva et.al., and L. Velho et.al., [9] introduced the field of non-Euclidean geometries especially hyperbolic geometry unfolded after its initial discovery. Research in the early stages functioned to prove that hyperbolic geometry stands alone as a valid structure despite existing independently from Euclidean foundations. Researches developed several models to study and visualize hyperbolic spaces during which the Poincaré disk model received specific attention together with the Poincaré half-plane model and the hyperboloid model. The separate models of hyperbolic space contained unique features which made them ideal for different aspects of analysis such as angle preservation, conformity and computational simplicity. In 2020 S. Huckemann et.al. and B. Eltzner et.al. [12] suggested the researchers studied fundamental structures in hyperbolic geometry that included hyperbolic lines and triangles as well as circles and tessellations. Doctors studied in detail the transformation of traditional Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 31 No. 4s (2024) 675 https://internationalpubls.com Euclidean distance measures together with angle summations and parallel concepts in hyperbolic systems. Through specific studies of hyperbolic trigonometry scientists gained knowledge needed to create formulas and metrics that excel in negatively curved spaces. Mathematical foundation that was developed allowed following generations to create new discoveries for theoretical and applied research. Scientists utilize hyperbolic geometry today to model complex networks through which hyperbolic space enables efficient data structure embedding despite its hierarchical and expansive aspect. Studies proved that hyperbolic models consume less space than Euclidean embeddings when representing internet topology social networks and biological systems because they effectively represent relationships. Many visualization techniques evolve to better show hyperbolic relationships and computer programs now let us explore dynamic views of geodesics and growth processes and tilings within hyperbolic spaces. In 2024 J. Miller et.al., D. Bhatia et.al., and S. Kobourov et.al., [5] proposed theoretical physics makes regular use of hyperbolic geometry mainly when studying general relativity together with negative-curvature cosmological systems. This mathematical concept has shaped new procedures in optimization as well as artistic developments of infinite structures through hyperbolic tessellations. Current education systems face obstacles to teach hyperbolic concepts intuitively to persons outside the pure mathematics field. The relevance in our time requires us to establish new educational tools and computational frameworks with interactive visualizations to help understand hyperbolic concepts better. Research on hyperbolic geometry demonstrates its fundamental nature as theoretical component and useful tool through ongoing efforts to increase knowledge and remove barriers for different fields of study. III. PROPOSED METHODOLOGY In order to deeply explore hyperbolic geometry and its properties, we begin by constructing a systematic mathematical framework that uses established models and introduces new ways to compute, visualize, and compare hyperbolic structures. The methodology is divided into five main phases: Model Selection, Geometric Definition, Metric Derivation, Visualization, and Application Testing [8]. A. Model Selection First, we select the Poincaré disk model as our primary working model due to its conformality, which preserves angles and makes calculations involving angles more intuitive. In the Poincaré disk, the set of points (𝑥, 𝑦) satisfying 𝑥2 + 𝑦2 < 1 defines the space. The metric in the disk is not Euclidean. Instead, the infinitesimal hyperbolic distance 𝑑𝑠 is given by: Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 31 No. 4s (2024) 676 https://internationalpubls.com 𝑑𝑠 = 2√𝑑𝑥2 + 𝑑𝑦2 1 − (𝑥2 + 𝑦2) This equation will serve as the foundation for further geometric derivations. B. Geometric Definition We define a hyperbolic line in the Poincaré disk as either a Euclidean circle orthogonal to the boundary or a diameter. The general form of a geodesic (line) is captured as: (𝑥 − 𝑎)2 + (𝑦 − 𝑏)2 = 𝑟2 where 𝑟2 = 𝑎2 + 𝑏2 − 1, ensuring orthogonality to the unit circle boundary. For two points 𝑃 and 𝑄 inside the disk, the hyperbolic distance 𝑑(𝑃, 𝑄) is computed as: 𝑑(𝑃, 𝑄) = arccosh (1 + 2‖𝑃 − 𝑄‖2 (1 − ‖𝑃‖2)(1 − ‖𝑄‖2) ) where || • || denotes the Euclidean norm. C. Metric Derivation A critical aspect of hyperbolic geometry is the calculation of areas and angles. In hyperbolic triangles, the sum of angles 𝛼 + 𝛽 + 𝛾 satisfies: 𝛼 + 𝛽 + 𝛾 < 𝜋 The area 𝑨 of a hyperbolic triangle is directly related to its angle deficit: 𝐴 = 𝜋 − (𝛼 + 𝛽 + 𝛾) Additionally, the hyperbolic law of cosines differs from its Euclidean counterpart. For a triangle with sides 𝑎, 𝑏, 𝑐 and opposite angles 𝛼, 𝛽, 𝛾, it is: cosh⁡(𝑐) = cosh⁡(𝑎)cosh⁡(𝑏) − sinh⁡(𝑎)sinh⁡(𝑏)cos⁡(𝛾) where cosh and sinh are hyperbolic cosine and sine functions, respectively. For circles in hyperbolic space, the circumference 𝐶 grows exponentially with radius 𝑟, and is given by: 𝐶 = 2𝜋sinh⁡(𝑟) rather than linearly, as in Euclidean geometry. The area 𝐴 enclosed by a hyperbolic circle of radius 𝑟 is: Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 31 No. 4s (2024) 677 https://internationalpubls.com 𝐴 = 2𝜋(cosh⁡(𝑟) − 1) thus demonstrating a strong departure from flat-space intuitions. D. Visualization To aid understanding, we implement computational simulations of hyperbolic structures using visualization libraries. Geodesics are drawn as arcs of circles, and hyperbolic tessellations are created using recursive symmetry operations. Given a transformation matrix 𝑇 for hyperbolic motions: 𝑇 = [ cosh⁡(𝜃) sinh⁡(𝜃) sinh⁡(𝜃) cosh⁡(𝜃) ] we simulate translations along geodesics. The transformations are essential for dynamically generating hyperbolic tilings and animations. Reflections across geodesics are another important operation, expressed mathematically by: 𝑃′ = 𝑅(𝑃) where 𝑅 is the reflection operator corresponding to a given geodesic. E. Application Testing Finally, we apply the methodology to simulate network embeddings into hyperbolic space. By minimizing the loss function: ℒ = ∑   𝑖,𝑗 𝑤𝑖𝑗(𝑑𝐻(𝑥𝑖, 𝑥𝑗) − 𝛿𝑖𝑗) 2 where 𝑤𝑖𝑗 are weights, 𝑑𝐻 denotes hyperbolic distance, and 𝛿𝑖𝑗 is the target distance, we optimize embeddings to preserve structural relationships. For optimization, we rely on hyperbolic gradient descent, where the update rule for a point 𝑥 is: 𝑥′ = exp𝑥⁡(−𝜂∇ℒ(𝑥)) with exp𝑥 being the exponential map at 𝑥, 𝜂 the learning rate, and ∇ℒ the Riemannian gradient. Throughout the process, continuous monitoring of error metrics such as Mean Squared Error (MSE) is maintained, defined as: MSE = 1 𝑛 ∑   𝑛 𝑖=1 (𝑑𝐻(𝑥𝑖, 𝑦𝑖) − 𝛿𝑖) 2 Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 31 No. 4s (2024) 678 https://internationalpubls.com to ensure the fidelity of the hyperbolic embedding’s. FIGURE 1: PROPOSED METHODOLOGY FOR HYPERBOLIC GEOMETRY STUDY IV. RESULT & DISCUSSIONS The deployment of this proposed method produced many valuable observations about hyperbolic spaces. Observing geodesics together with tessellations under the Poincaré disk model helped demonstrate mathematically as well as visually how angular measures stayed intact while distance measurements were altered. Spatial intuition works substantially different in hyperbolic space because distances become dramatically greater near the boundary. Calculation results showed that randomly picked hyperbolic distance measurements exceeded their equivalent Euclidean values in all tests thereby validating theoretical hyperbolic expansion principles [10]. Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 31 No. 4s (2024) 679 https://internationalpubls.com Area expansion between Euclidean and hyperbolic circles was evaluated as one method to analyze structural properties. The hyperbolic space shows faster growth of area compared to Euclidean space when both spaces experience a similar increase in radius (Figure 2 illustrates the results). The non-linear area growth demonstrates both the fundamental curvature principles while showing that basic geometric structures have totally different behaviors. Boundary limitations enable hyperbolic geometries to contain substantially more space than flat geometries because of their distinctive circle growth patterns. FIGURE 2: AREA GROWTH COMPARISON BETWEEN EUCLIDEAN AND HYPERBOLIC CIRCLES The research also focused on developing tessellation examples through regular polygons in hyperbolic space. The analysis examined how triangles and pentagons tessellate side by side to bring out angular deficits that create unique patterns that cannot occur in Euclidean geometry. The hyperbolic geometry showed a greater number of connecting polygons at each singular point than occurs in Euclidean tilings according to Table 1. TABLE 1: NUMBER OF POLYGONS AT A SINGLE VERTEX IN DIFFERENT GEOMETRIES Geometry Type Polygon Type Number of Polygons at Vertex Euclidean Triangle 6 Hyperbolic Triangle 7 or more Euclidean Pentagon Impossible Hyperbolic Pentagon 4 or 5 3.14 12.57 28.27 3.88 20.31 59.52 0 10 20 30 40 50 60 70 1 2 3 Area Growth Comparison between Euclidean and Hyperbolic Circles Euclidean Area (πr²) Hyperbolic Area (Approx.) Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 31 No. 4s (2024) 680 https://internationalpubls.com The evaluation of large datasets through hyperbolic embedding required simulated network graphs to be embedded into Euclidean and hyperbolic spaces. Table 2 presents the summary of performance metrics regarding average distortion together with path preservation results (see Table 2). TABLE 2: PERFORMANCE METRICS FOR NETWORK EMBEDDINGS IN EUCLIDEAN VS HYPERBOLIC SPACE Metric Euclidean Embedding Hyperbolic Embedding Average Distortion 0.42 0.15 Shortest Path Preservation 73% 89% Average Embedding Error 0.39 0.21 The visual representation of embedding’s became vital to observe the spatial structures within the data. The networks in Euclidean space experienced stretching while needing substantial amounts of space to preserve their network integrity. The natural structure of hyperbolic embedding’s could compact networks effectively since they preserved proximity and hierarchy seamlessly. The two embedding’s possess different structural characteristics which show up in the network graph visualizations (see Figure 3). FIGURE 3: NETWORK EMBEDDING COMPARISON IN EUCLIDEAN VS HYPERBOLIC SPACES The tests demonstrated that connections between nodes in hyperbolic geometry featured easier routes which required less complex paths than Euclidean geometry would have needed. Experimental results from all investigations prove that natural hyperbolic structures offer superior data representation capabilities for extensive hierarchical information systems. 5.2 7.8 6.4 3.1 4.5 3.8 0 2 4 6 8 10 A - B B - C C - D Network Embedding Comparison in Euclidean vs Hyperbolic Spaces Euclidean Distance (units) Hyperbolic Distance (units) Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 31 No. 4s (2024) 681 https://internationalpubls.com Test participants needed time to adjust their understanding of hyperbolic visualization rules but eventually showed quick adoption of these rules during their interactions. The general usefulness of hyperbolic space visualizations becomes possible in real applications involving virtual reality environments and educational software and data science tools because of their spatial efficiency and visual hierarchy capabilities. The research verified theory-based predictions in addition to creating practical potential for hyperbolic implementation across modern technology fields. The combination of visual and computational and comparative evidence demonstrates that hyperbolic geometry serves as an advanced tool beyond academic theories because it delivers effective solutions for current science and engineering challenges. V. CONCLUSION The alternative system of hyperbolic geometry substitutes the traditional concepts of Euclidean intuitions to create new frameworks of understanding geometric space and spatial relationships. Theoretical along with computational studies brought forward diverse behavioral patterns which exist within hyperbolic spaces and their corresponding mathematical models. Modern scientific interrogation relies on hyperbolic geometry for extensive physical applications and computing applications together with artistic applications. REFERENCES [1] N. Abrosimov and A. Mednykh, “Area and volume in non-Euclidean geometry,” in EMS Press eBooks, 2019, pp. 151–189. doi: 10.4171/196-1/11. [2] D. Ferrarello, M. F. Mammana, and E. Taranto, “Non-Euclidean Geometry with Art by Means of GeoGebra,” International Journal for Technology in Mathematics Education, vol. 26, no. 3, pp. 113–119, Sep. 2019, doi: 10.1564/tme_v26.3.02. [3] I. Stevenson, “Hyperbolic Space: Designing a Computational Context for Learning Non-Euclidean Geometry,” International Journal of Computers for Mathematical Learning, vol. 5, no. 2, pp. 143–167, Jan. 2000, doi: 10.1023/a:1009802004645. [4] N. A’Campo and A. Papadopoulos, “On Klein’s So-called Non-Euclidean geometry,” in EMS Press eBooks, 2015, pp. 91–136. doi: 10.4171/148-1/5. [5] J. Miller, D. Bhatia, and S. Kobourov, “State of the art of graph visualization in non‐ Euclidean spaces,” Computer Graphics Forum, vol. 43, no. 3, Jun. 2024, doi: 10.1111/cgf.15113. [6] C. Saxena, T. Liu, and I. King, “A survey of graph curvature and embedding in Non- Euclidean spaces,” in Lecture notes in computer science, 2020, pp. 127–139. doi: 10.1007/978-3-030-63833-7_11. [7] L. Moreno-Armella, C. Brady, and R. Elizondo-Ramirez, “Dynamic hyperbolic geometry: building intuition and understanding mediated by a Euclidean model,” International Journal of Mathematical Education in Science and Technology, vol. 49, no. 4, pp. 594–612, Jan. 2018, doi: 10.1080/0020739x.2017.1418915. Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 31 No. 4s (2024) 682 https://internationalpubls.com [8] S. D. Casey, “Harmonic Analysis in Non-Euclidean Spaces: Theory and Application,” in Applied and numerical harmonic analysis, 2017, pp. 565–601. doi: 10.1007/978-3- 319-55556-0_6. [9] T. Novello, V. Da Silva, and L. Velho, “Global illumination of non-Euclidean spaces,” Computers & Graphics, vol. 93, pp. 61–70, Oct. 2020, doi: 10.1016/j.cag.2020.09.014. [10] S. Sanborn et al., “Beyond Euclid: An Illustrated Guide to Modern Machine Learning with Geometric, Topological, and Algebraic Structures,” arXiv (Cornell University), Jul. 2024, doi: 10.48550/arxiv.2407.09468. [11] J. Batle, A. Bagdasaryan, M. Abdel-Aty, and S. Abdalla, “Generalized Thomson problem in arbitrary dimensions and non-euclidean geometries,” Physica a Statistical Mechanics and Its Applications, vol. 451, pp. 237–250, Feb. 2016, doi: 10.1016/j.physa.2016.01.069. [12] S. Huckemann and B. Eltzner, “Statistical methods Generalizing principal component analysis to Non-Euclidean spaces,” in Springer eBooks, 2020, pp. 317–338. doi: 10.1007/978-3-030-31351-7_10. [13] W. Peng, T. Varanka, A. Mostafa, H. Shi, and G. Zhao, “Hyperbolic Deep Neural Networks: a survey,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 44, no. 12, pp. 10023–10044, Dec. 2021, doi: 10.1109/tpami.2021.3136921. [14] Y. L. Sukestiyarno, K. U. Z. Nugroho, S. Sugiman, and B. Waluya, “Learning trajectory of non-Euclidean geometry through ethnomathematics learning approaches to improve spatial ability,” Eurasia Journal of Mathematics Science and Technology Education, vol. 19, no. 6, p. em2285, May 2023, doi: 10.29333/ejmste/13269. [15] A. A. Ungar, “Gyrations: The Missing Link Between Classical Mechanics with Its Underlying Euclidean Geometry and Relativistic Mechanics with Its Underlying Hyperbolic Geometry,” in Springer eBooks, 2012, pp. 463–504. doi: 10.1007/978-3- 642-28821-0_18.