In dieser Arbeit präsentieren wir eine neue Art von Newton-Verfahren mit Liniensuche, basierend auf Interpolation im Bildbereich nach Wedin et al. [LW84]. Von dem resultierenden stabilisierten Newton-Algorithmus wird theoretisch und praktisch gezeigt, dass er effizient ist im Falle von nichtsingulären Lösungen. Darüber hinaus wird beobachtet, dass er eine superlineare Rate von Konvergenz bei einfachen Singularitäten erhält. Hingegen ist vom Newton-Verfahren ohne Liniensuche bekannt, dass es nur linear von fast allen Punkten in der Nähe einer singulären Lösung konvergiert. In Hinsicht auf Anwendungen auf Komplementaritätsprobleme betrachten wir auch Systeme, deren Jacobimatrix nicht differenzierbar sondern nur semismooth ist. Auch hier erreicht unser stabilisiertes und beschleunigtes Newton- Verfahren Superlinearität bei einfachen Singularitäten. / In this thesis we present a new type of line-search for Newton’s method, based on range space interpolation as suggested by Wedin et al. [LW84]. The resulting stabilized Newton algorithm is theoretically and practically shown to be efficient in the case of nonsingular roots. Moreover it is observed that it maintains a superlinear rate of convergence at simple singularities. Whereas Newton’s method without line-search is known to converge only linearly from almost all points near the singular root. In view of applications to complementarity problems we also consider systems, whose Jacobian is not differentiable but only semismooth. Again, our stabilized and accelerated Newton’s method achieves superlinearity at simple singularities.

Όπως φαίνεται και από τον τίτλο της, ο σκοπός της Διπλωματικής αυτής Εργασίας είναι διπλός. Αφ’ ενός διαπραγματεύεται ένα κλασικό μαθηματικό πρόβλημα, το πρόβλημα Riemann-Hilbert (RH), που παρουσιάζεται και επιλύεται σε μια σειρά περιπτώσεων. Αφ’ ετέρου παρουσιάζεται η εφαρμογή του προβλήματος αυτού στη μελέτη προβλημάτων αρχικών ή αρχικών-συνοριακών τιμών για γραμμικές και μη γραμμικές μερικές διαφορικές εξισώσεις. Η εργασία διαρθρώνεται σε τεσσερα (4) κεφάλαια. Ακριβέστερα, η δομή των κεφαλαίων είναι η ακόλουθη. Το πρώτο κεφάλαιο αποτελεί την εισαγωγή της εργασίας και περιέχει, εκτός από μια εποπτική παρουσίαση του προβλήματος, μια σύντομη ιστορική αναδρομή καθώς και παράθεση των εφαρμογών του προβλήματος. Το δεύτερο κεφάλαιο τιτλοφορείται ‘Ολοκληρώματα τύπου Cauchy’ και είναι αφιερωμένο στην παρουσίαση του αναγκαίου υποβάθρου, με σκοπό να είναι η ακόλουθη παρουσίαση αυτάρκης. Τα θέματα που διαπραγματεύεται είναι: Oλοκληρώματα τύπου Cauchy, συναρτήσεις τύπου Hölder, ολοκληρώματα κύριας τιμής του Cauchy, θεώρημα των Plemelj-Sokhotski, ολοκληρωτικός τελεστής του Cauchy, ολοκληρώματα τύπου Cauchy στην πραγματική ευθεία. Το τρίτο κεφάλαιο, ‘Το πρόβλημα Riemann-Hilbert’, παρουσιάζει το πρόβλημα καθώς και την επίλυσή του σε μια σειρά περιπτώσεων. Στην πιο απλή διατύπωσή του, το πρόβλημα ζητά τον προσδιορισμό μιας τμηματικά ολόμορφης μιγαδικής συνάρτησης μιας μιγαδικής μεταβλητής η οποία παρουσιάζει δοσμένο άλμα κατά μήκος δοσμένης καμπύλης του μιγαδικού επιπέδου. Εστιαζόμαστε αποκλειστικά σε βαθμωτά προβλήματα. Επίσης, εργαζόμαστε με συνοριακές καμπύλες που έχουν την ιδιότητα να χωρίζουν το μιγαδικό επίπεδο σε δύο τμήματα: κλειστές καμπύλες, καθώς και την πραγματική ευθεία. Ειδικότερα, αναλύονται τα ακόλουθα προβλήματα: (i) Πρόβλημα Riemann-Hilbert (RH) για κλειστές καμπύλες: (1) Aθροιστικό (additive) πρόβλημα RH. (2) Πρόβλημα παραγοντοποίησης (factorization) RH. (3) Γενικό μη ομογενές πρόβλημα RH. (ii) Πρόβλημα RH επί της πραγματικής ευθείας: (1) Aθροιστικό (additive) πρόβλημα RH. (2) Πρόβλημα παραγοντοποίησης (factorization) RH. (3) Γενικό μη ομογενές πρόβλημα RH. Το τέταρτο κεφάλαιο τιτλοφορείται ‘Προβλήματα Αρχικών-Συνοριακών Τιμών για Γραμμικές και μη Γραμμικές Μερικές Διαφορικές Εξισώσεις’. Εδώ διαπραγματευόμαστε μερικές διαφορικές εξισώσεις (ΜΔΕ), τόσο γραμμικές όσο και μη γραμμικές, που έχουν την ιδιότητα να διαθέτουν ζεύγος Lax (Lax pair formulation): Aυτό σημαίνει ότι κάθε μία από αυτές τις ΜΔΕ μπορεί να γραφεί σαν η συνθήκη συμβατότητας (ολοκληρωσιμότητας) ενός ζεύγους γραμμικών ΜΔΕ, που περιέχει και μια ελεύθερη μιγαδική παράμετρο (φασματική παράμετρος). Τέτοιες ΜΔΕ χαρακτηρίζονται και σαν ολοκληρώσιμες (integrable) με τη μέθοδο της αντίστροφης σκέδασης (inverse scattering method). Η τελευταία αποτελεί μια μέθοδο επίλυσης του προβλήματος αρχικών τιμών, ή Cauchy, για εξελικτικές ΜΔΕ αυτού του είδους. Η νεότερη μέθοδος του ενοποιημένου φασματικού μετασχηματισμού (unified transform method), ή της ταυτόχρονης φασματικής ανάλυσης (simultaneous spectral analysis) του ζεύγους Lax, γενικεύει την προηγούμενη μέθοδο με τρόπο που να μπορεί να εφαρμοστεί και σε προβλήματα αρχικών-συνοριακών τιμών τέτοιων ΜΔΕ (και όχι μόνο). Στο κεφάλαιο αυτό της εργασίας μελετιούνται τα ακόλουθα προβλήματα. (i). Το πρόβλημα αρχικών τιμών (ΠΑΤ) για τη (γραμμική) ΜΔΕ της διάχυσης (ή θερμότητας) (heat (or diffusion) equation). Εδώ παρουσιάζεται η μέθοδος της αντίστροφης σκέδασης στην απλούστερή της μορφή. (ii). Ένα αρκετά γενικό φασματικό πρόβλημα, που μπορεί να αποτελέσει το χωρικό μέρος του ζευγαριού Lax για μια πλειάδα μη γραμμικών ΜΔΕ. Στη συνέχεια, η προσοχή μας εστιάζεται στο λεγόμενο φασματικό πρόβλημα των Zakharov-Shabat. Σαν εφαρμογή, μελετάται το ΠΑΤ για τη μη γραμμική Εξίσωση Schrodinger (Nonlinear Schrodinger, NLS). (iii). Το πρόβλημα αρχικών-συνοριακών τιμών (ΠΑΣΤ) για την εξίσωση της διάχυσης ορισμένη στην ημιευθεία της χωρικής μεταβλητής. Εδώ περιγράφεται η μέθοδος του ενοποιημένου φασματικού μετασχηματισμού στην απλούστερή της μορφή, εφαρμοζόμενη δηλαδή σε ένα γραμμικό πρόβλημα. H εργασία καταλήγει με την παράθεση της βιβλιογραφίας, σύμφωνα με τις αναφορές που προκύπτουν από το κείμενο. / As it is shown in its title, the purpose of this M.Sc.thesis is twofold. First, we discuss a classical mathematical problem, called the Riemann-Hilbert problem. This problem is presented and solved in a series of cases. Afterwards, we present the applications of this problem to the study of initial value problems and initial-boundary value problems for linear and nonlinear partial differential equations. The thesis is organized in four (4) chapters. More accurately, the structure of the four chapters is as follows. The first chapter constitutes of the Introduction to the thesis. It contains the presentation of the problem, a short historical retrospection of the problem, as well as a list of applications of the problem. The second chapter, entitled “Cauchy Type Integrals”, is dedicated to the presentation of the necessary background, so as to make the following presentation self-contained. The topics negotiated are: Cauchy type integrals, Hölder type functions, Cauchy principal value integrals, the Plemelj-Sokhotski theorem, the Cauchy integral operator, Cauchy type integrals on the real line. The third chapter, “The Riemann-Hilbert Problem”, presents the problem, as well s its solution, in a series of cases. The problem’s simplest formulation seeks for a sectionally holomorphic, complex valued function of a single complex variable, which undergoes a given (predetermined) jump along a given curve of the complex plane. We focus our attention exclusively on scalar Riemann-Hilbert problems. We work exclusively with discontinuity curves that have the property to divide the complex plane into two sections, and, in particular, with closed curves, as well as with the real line. In particular, we analyse the following problems: (i). The Riemann-Hilbert (RH) problem for closed curves: (1). Additive RH problem. (2). Factorization RH problem. (3). General non-homogeneous RH problem. (ii). RH problem on the real line. (1). Additive RH problem. (2). Factorization RH problem. (3). General non-homogeneous RH problem. The fourth chapter is entitled “Initial-Boundary Value Problems for Linear and Nonlinear Partial Differential Equations”. Here we negotiate with patial differential equations (PDE), linear as well as nolinear, which have the distinguishing property of possessing a so-called Lax pair formulation. By this we mean that, any of these PDEs is equivalent to the compatibility (integrability) condition of a proper pair of linear differential equations, the so-called Lax pair, that also contains a free complex parameter, termed to the spectral parameter. Such PDEs are also characterized as integrable by the inverse scattering method. The last method, also called the inverse spectral method, is a method for solving the initial value problem, or Cauchy problem, for evolutionary PDEs of this kind. The new method of simultaneous spectral analysis of the Lax pair, also called the unified transform method, generalizes the previous one in a manner that renders it applicable also to initial-boundary value problems for such PDEs. In this, fourth, chapter we study the following problems: (i). The initial value problem for the (linear) heat (or diffusion) equation. Here is presented the inverse scattering method in its simplest form. (ii). An adequately general spectral problem, which may constitute the spatial part of the Lax pair for many integrable nonlinear PDEs. We afterwards focus our attention to a specific case of this problem, the so-called Zakharov-Shabat spectral problem. As an application, we study the initial value problem for the so-called Nonlinear Schrodinger (NLS) equation. (iii). The initial-boundary value problem for the heat (or diffusion) equation posed on a semi-infinite interval of the spatial variable. Here we present the unified transform method in its simplest form, i.e., applied on a linear problem. The thesis terminates with the presentation of the bibliography, in accordance with the references that appear in the text.

[pt] Primeiro examinamos soluções de viscosidade Lp para equações elípticas totalmente não lineares com ingredientes de fronteira mensuráveis. Ao considerar p0 < p < d, focamos nas estimativas da regularidade dos gradientes derivadas de potenciais não lineares. Encontramos condições para Lipschitz-continuidade local das soluções e continuidade do gradiente. Examinamos avanços recentes na teoria da regularidade decorrentes de estimativas potenciais (não lineares). Nossas descobertas decorrem de – e são inspiradas por – fatos fundamentais na teoria de soluções de Lp-viscosidade, e resultados do trabalho de Panagiota Daskalopoulos, Tuomo Kuusi e Giuseppe Mingione (DKM2014). Na segunda parte provamos a regularidade parcial de mapas harmônicos com peso fracamente estacionários com dados de fronteira livre em um cone. Como ponto de partida, damos uma olhada na teoria da regularidade parcial interior para mapas harmônicos fracionários de minimização de energia intrínseca do espaço euclidiano em variedades Riemannianas compactas e suaves para potências fracionárias estritamente entre zero e um. Mapas harmônicos fracionários intrínsecos podem ser estendidos para mapas harmônicos com peso, então provamos regularidade parcial para mapas harmônicos minimizantes locais com dados de fronteira (parcialmente) livres em meios-espaços, mapas harmônicos fracionários então herdam essa regularidade. / [en] We first examine Lp-viscosity solutions to fully nonlinear elliptic equations with bounded measurable ingredients. By considering p0 < p < d, we focus on gradient-regularity estimates stemming from nonlinear potentials. We find conditions for local Lipschitz-continuity of the solutions and continuity of the gradient. We survey recent breakthroughs in regularity theory arising from (nonlinear) potential estimates. Our findings follow from – and are inspired by – fundamental facts in the theory of Lp-viscosity solutions, and results in the work of Panagiota Daskalopoulos, Tuomo Kuusi and Giuseppe Mingione (DKM2014). In the second part we prove partial regularity of weakly stationary weighted harmonic maps with free boundary data on a cone. As a starting point we take a look at the interior partial regularity theory for intrinsic energy minimising fractional harmonic maps from Euclidean space into smooth compact Riemannian manifolds for fractional powers strictly between zero and one. Intrinsic fractional harmonic maps can be extended to weighted harmonic maps, so we prove partial regularity for locally minimising harmonic maps with (partially) free boundary data on half-spaces, fractional harmonic maps then inherit this regularity.

This thesis presents a mathematical model of the nanoporous anode within a dyesensitised solar cell (DSC). The main purpose of this work is to investigate interfacial charge transfer and charge transport within the porous anode of the DSC under both illuminated and non-illuminated conditions. Within the porous anode we consider many of the charge transfer reactions associated with the electrolyte species, adsorbed dye molecules and semiconductor electrons at the semiconductor-dye- electrolyte interface. Each reaction at this interface is modelled explicitly via an electrochemical equation, resulting in an interfacial model that consists of a coupled system of non-linear algebraic equations. We develop a general model framework for charge transfer at the semiconductor-dye-electrolyte interface and simplify this framework to produce a model based on the available interfacial kinetic data. We account for the charge transport mechanisms within the porous semiconductor and the electrolyte filled pores that constitute the anode of the DSC, through a one- dimensional model developed under steady-state conditions. The governing transport equations account for the diffusion and migration of charge species within the porous anode. The transport model consists of a coupled system of non-linear differential equations, and is coupled to the interfacial model via reaction terms within the mass-flux balance equations. An equivalent circuit model is developed to account for those components of the DSC not explicitly included in the mathematical model of the anode. To obtain solutions for our DSC mathematical model we develop code in FORTRAN for the numerical simulation of the governing equations. We additionally employ regular perturbation analysis to obtain analytic approximations to the solutions of the interfacial charge transfer model. These approximations facilitate a reduction in computation time for the coupled mathematical model with no significant loss of accuracy. To obtain predictions of the current generated by the cell we source kinetic and transport parameter values from the literature and from experimental measurements associated with the DSC commissioned for this study. The model solutions we obtain with these values correspond very favourably with experimental data measured from standard DSC configurations consisting of titanium dioxide porous films with iodide/triiodide redox couples within the electrolyte. The mathematical model within this thesis enables thorough investigation of the interfacial reactions and charge transport within the DSC.We investigate the effects of modified cell configurations on the efficiency of the cell by varying associated parameter values in our model. We find, given our model and the DSC configuration investigated, that the efficiency of the DSC is improved with increasing electron diffusion, decreasing internal resistances and with decreasing dark current. We conclude that transport within the electrolyte, as described by the model, appears to have no limiting effect on the current predicted by the model until large positive voltages. Additionally, we observe that the ultrafast injection from the excited dye molecules limits the interfacial reactions that affect the DSC current.

Tesis por compendio / [ES] La resolución de ecuaciones y sistemas de ecuaciones no lineales es fundamental en muchas disciplinas científicas y de ingeniería, incluyendo la física, la química, la biología, la economía y la informática. Los métodos numéricos son cruciales para resolver estas ecuaciones debido a su complejidad, que a menudo resulta en múltiples soluciones o en la ausencia de ellas, lo que hace que los métodos analíticos tradicionales sean inadecuados. Esta investigación se centra en el desarrollo y análisis de nuevos esquemas iterativos para resolver ecuaciones y sistemas de ecuaciones no lineales, enfatizando la convergencia, la estabilidad y la eficiencia computacional. Como parte de esta investigación se publicaron tres artículos clave. El primer artículo introduce una novedosa familia de métodos iterativos de dos pasos derivada de un esquema de Newton amortiguado, que incluye un paso adicional de Newton con una función de peso y una derivada "congelada". Esta familia, inicialmente una clase de cuatro parámetros con convergencia de primer orden, se convierte en una familia de un solo parámetro con convergencia de tercer orden, que además muestra una estabilidad y eficiencia excepcionales, validadas mediante pruebas numéricas. El segundo artículo presenta un nuevo método iterativo de tres pasos, inicialmente una familia de tres parámetros de cuarto orden que acelera a una familia de un solo parámetro de sexto orden. La convergencia, la dinámica compleja y el comportamiento numérico de este método son estudiados a fondo, identificando miembros estables adecuados para problemas prácticos. El tercer artículo extiende la familia de sexto orden a sistemas de ecuaciones no lineales, creando un esquema de un solo parámetro altamente eficiente. Los análisis dinámicos y numéricos confirman la convergencia, estabilidad y aplicabilidad de esta familia extendida para problemas de gran escala. La investigación tiene como objetivo superar las limitaciones de algunos métodos existentes, ofreciendo soluciones robustas y eficientes para ecuaciones y sistemas no lineales. El documento está estructurado para cubrir el desarrollo, análisis y validación de estos métodos, proporcionando recomendaciones específicas para su aplicación práctica en varios dominios científicos y de ingeniería. / [CA] La resolució d'equacions i sistemes d'equacions no lineals és fonamental en moltes disciplines científiques i d'enginyeria, incloent la física, la química, la biologia, l'economia i la informàtica. Els mètodes numèrics són crucials per a resoldre aquestes equacions a causa de la seua complexitat, que sovint resulta en múltiples solucions o en l'absència d'elles, la qual cosa fa que els mètodes analítics tradicionals siguen inadequats. Aquesta investigació se centra en el desenvolupament i anàlisi de nous esquemes iteratius per a resoldre equacions i sistemes d'equacions no lineals, emfatitzant la convergència, l'estabilitat i l'eficiència computacional. Com a part d'aquesta investigació es van publicar tres articles clau. El primer article introdueix una nova família de mètodes iteratius de dos passos derivada d'un esquema de Newton esmorteït, que inclou un pas addicional de Newton amb una funció de pes i una derivada "congelada". Aquesta família, inicialment una classe de quatre paràmetres amb convergència de primer ordre, es converteix en una família d'un sol paràmetre amb convergència de tercer ordre, que a més mostra una estabilitat i eficiència excepcionals, validats mitjançant proves numèriques. El segon article presenta un nou mètode iteratiu de tres passos, inicialment una família de tres paràmetres de quart ordre que accelera a una família d'un sol paràmetre de sisè ordre. La convergència, la dinàmica complexa i el comportament numèric d'aquest mètode són estudiats a fons, identificant membres estables adequats per a problemes pràctics. El tercer article amplia la família de sisè ordre a sistemes d'equacions no lineals, creant un esquema d'un sol paràmetre altament eficient. Els anàlisis dinàmics i numèrics confirmen la convergència, estabilitat i aplicabilitat d'aquesta família ampliada per a problemes de gran escala. La investigació té com a objectiu superar les limitacions d'alguns mètodes existents, oferint solucions robustes i eficients per a equacions i sistemes no lineals. El document està estructurat per a cobrir el desenvolupament, anàlisi i validació d'aquests mètodes, proporcionant recomanacions específiques per a la seua aplicació pràctica en diversos dominis científics i d'enginyeria. / [EN] The resolution of non-linear equations and systems is fundamental in various scientific and engineering fields, including physics, chemistry, biology, economics, and computer science. Numerical methods are crucial for solving these equations due to their complexity, which often results in multiple or no solutions, rendering traditional analytical methods inadequate. This research focuses on developing and analyzing new iterative schemes for solving non-linear equations and systems, emphasizing convergence, stability, and computational efficiency. Three key papers were published as part of this research. The first paper introduces a novel family of two-step iterative methods derived from a damped Newton scheme, which includes an additional Newton step with a weight function and a "frozen" derivative. This family, initially a four-parameter class with first-order convergence, becomes a single-parameter family with third-order convergence, which also exhibits exceptional stability and efficiency, validated through numerical tests. The second paper presents a new three-step iterative method, initially a three-parameter fourth-order family, which accelerates to a single-parameter sixth-order family. This method's convergence, complex dynamics, and numerical behavior are thoroughly studied, identifying stable members suitable for practical problems. The third paper extends the sixth-order family to systems of non-linear equations, creating a highly efficient single-parameter family. Dynamic and numerical analyses confirm the convergence, stability, and applicability of this extended family for large-scale problems. The research aims to overcome the limitations of some existing methods, offering robust and efficient solutions for non-linear equations and systems. The document is structured to cover the development, analysis, and validation of these methods, providing specific recommendations for their practical application in various scientific and engineering domains. / Moscoso Martínez, ME. (2024). Efficient Numerical Methods for Solving Nonlinear Problems [Tesis doctoral]. Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/212946 / Compendio

No description available.