Dissertations / Theses on the topic 'Fredholm theory'

Thesis (MSc (Mathematics))--University of Stellenbosch, 2010.
ENGLISH ABSTRACT: This thesis is a study of a generalisation, due to R. Harte (see [9]), of Fredholm theory in the context of bounded linear operators on Banach spaces to a theory in a Banach algebra setting. A bounded linear operator T on a Banach space X is Fredholm if it has closed range and the dimension of its kernel as well as the dimension of the quotient space X/T(X) are finite. The index of a Fredholm operator is the integer dim T−1(0)−dimX/T(X). Weyl operators are those Fredholm operators of which the index is zero. Browder operators are Fredholm operators with finite ascent and descent. Harte’s generalisation is motivated by Atkinson’s theorem, according to which a bounded linear operator on a Banach space is Fredholm if and only if its coset is invertible in the Banach algebra L(X) /K(X), where L(X) is the Banach algebra of bounded linear operators on X and K(X) the two-sided ideal of compact linear operators in L(X). By Harte’s definition, an element a of a Banach algebra A is Fredholm relative to a Banach algebra homomorphism T : A ! B if Ta is invertible in B. Furthermore, an element of the form a + b where a is invertible in A and b is in the kernel of T is called Weyl relative to T and if ab = ba as well, the element is called Browder. Harte consequently introduced spectra corresponding to the sets of Fredholm, Weyl and Browder elements, respectively. He obtained several interesting inclusion results of these sets and their spectra as well as some spectral mapping and inclusion results. We also convey a related result due to Harte which was obtained by using the exponential spectrum. We show what H. du T. Mouton and H. Raubenheimer found when they considered two homomorphisms. They also introduced Ruston and almost Ruston elements which led to an interesting result related to work by B. Aupetit. Finally, we introduce the notions of upper and lower semi-regularities – concepts due to V. M¨uller. M¨uller obtained spectral inclusion results for spectra corresponding to upper and lower semi-regularities. We could use them to recover certain spectral mapping and inclusion results obtained earlier in the thesis, and some could even be improved.
AFRIKAANSE OPSOMMING: Hierdie tesis is ‘n studie van ’n veralgemening deur R. Harte (sien [9]) van Fredholm-teorie in die konteks van begrensde lineˆere operatore op Banachruimtes tot ’n teorie in die konteks van Banach-algebras. ’n Begrensde lineˆere operator T op ’n Banach-ruimte X is Fredholm as sy waardeversameling geslote is en die dimensie van sy kern, sowel as di´e van die kwosi¨entruimte X/T(X), eindig is. Die indeks van ’n Fredholm-operator is die heelgetal dim T−1(0) − dimX/T(X). Weyl-operatore is daardie Fredholm-operatore waarvan die indeks gelyk is aan nul. Fredholm-operatore met eindige styging en daling word Browder-operatore genoem. Harte se veralgemening is gemotiveer deur Atkinson se stelling, waarvolgens ’n begrensde lineˆere operator op ’n Banach-ruimte Fredholm is as en slegs as sy neweklas inverteerbaar is in die Banach-algebra L(X) /K(X), waar L(X) die Banach-algebra van begrensde lineˆere operatore op X is en K(X) die twee-sydige ideaal van kompakte lineˆere operatore in L(X) is. Volgens Harte se definisie is ’n element a van ’n Banach-algebra A Fredholm relatief tot ’n Banach-algebrahomomorfisme T : A ! B as Ta inverteerbaar is in B. Verder word ’n Weyl-element relatief tot ’n Banach-algebrahomomorfisme T : A ! B gedefinieer as ’n element met die vorm a + b, waar a inverteerbaar in A is en b in die kern van T is. As ab = ba met a en b soos in die definisie van ’n Weyl-element, dan word die element Browder relatief tot T genoem. Harte het vervolgens spektra gedefinieer in ooreenstemming met die versamelings van Fredholm-, Weylen Browder-elemente, onderskeidelik. Hy het heelparty interessante resultate met betrekking tot insluitings van die verskillende versamelings en hulle spektra verkry, asook ’n paar spektrale-afbeeldingsresultate en spektraleinsluitingsresultate. Ons dra ook ’n verwante resultaat te danke aan Harte oor, wat verkry is deur van die eksponensi¨ele-spektrum gebruik te maak. Ons wys wat H. du T. Mouton en H. Raubenheimer verkry het deur twee homomorfismes gelyktydig te beskou. Hulle het ook Ruston- en byna Rustonelemente gedefinieer, wat tot ’n interessante resultaat, verwant aan werk van B. Aupetit, gelei het. Ten slotte stel ons nog twee begrippe bekend, naamlik ’n onder-semi-regulariteit en ’n bo-semi-regulariteit – konsepte te danke aan V. M¨uller. M¨uller het spektrale-insluitingsresultate verkry vir spektra wat ooreenstem met bo- en onder-semi-regulariteite. Ons kon dit gebruik om sekere spektrale-afbeeldingsresultate en spektrale-insluitingsresultate wat vroe¨er in hierdie tesis verkry is, te herwin, en sommige kon selfs verbeter word.

Doutoramento em Matemática
Na presente tese consideramos combinações algébricas de operadores de Wiener-Hopf e de Hankel com diferentes classes de símbolos de Fourier. Nomeadamente, foram considerados símbolos matriciais na classe de elementos quase periódicos, semi-quase periódicos, quase periódicos por troços e certas funções matriciais sectoriais. Adicionalmente, foi dedicada atenção também aos operadores de Toeplitz mais Hankel com símbolos quase periódicos por troços e com símbolos escalares possuindo n pontos de discontinuidades quase periódicas usuais. Em toda a tese, um objectivo principal teve a ver com a obtenção de descrições para propriedades de Fredholm para estas classes de operadores. De forma a deduzir a invertibilidade lateral ou bi-lateral para operadores de Wiener-Hopf mais Hankel com símbolos matriciais AP foi introduzida a noção de factorização assimétrica AP. Neste âmbito, foram dadas condições suficientes para a invertibilidade lateral e bi-lateral de operadores de Wiener- Hopf mais Hankel com símbolos matriciais AP. Para tais operadores, foram ainda exibidos inversos generalizados para todos os casos possíveis. Para os operadores de Wiener-Hopf-Hankel com símbolos matriciais SAP e PAP foi deduzida a propriedade de Fredholm e uma fórmula para a soma dos índices de Fredholm destes operadores de Wiener-Hopf mais Hankel e operadores de Wiener-Hopf menos Hankel. Uma versão mais forte destes resultados foi obtida usando a factorização generalizada AP à direita. Foram analisados os operadores de Wiener-Hopf-Hankel com símbolos que apresentam determinadas propriedades pares e também com símbolos de Fourier que contêm matrizes sectoriais. Em adição, para operadores de Wiener-Hopf-Hankel, foi obtido um resultado correspondente ao teorema clássico de Douglas e Sarason conhecido para operadores de Toeplitz com símbolos sectoriais e unitários. Condições necessárias e suficientes foram também deduzidas para que os operadores de Wiener-Hopf mais Hankel com símbolos L∞ sejam de Fredholm (ou invertíveis). Para se obter tal resultado, trabalhou-se com certas factorizações ímpares dos símbolos de Fourier. Os operadores de Toeplitz mais Hankel gerados por símbolos que possuem n pontos de discontinuidades quase periódicas usuais foram também considerados. Foram obtidas condições sob as quais estes operadores são invertíveis à direita e com dimensão de núcleo infinita, invertíveis à esquerda e com dimensão de co-núcleo infinita ou não normalmente solúveis. A nossa atenção foi também colocada em operadores de Toeplitz mais Hankel com símbolos matriciais contínuos por troços. Para tais operadores, condições necessárias e suficientes foram obtidas para se ter a propriedade de Fredholm. Tal foi realizado usando a abordagem do cálculo simbólico, determinados operadores auxiliares emparelhados com símbolos semi-quase periódicos e várias relações de equivalência após extensão entre operadores.
In this thesis we considered algebraic combinations of Wiener-Hopf and Hankel operators with different classes of Fourier symbols. Namely, matrix symbols from the almost periodic, semi-almost periodic, piecewise almost periodic and certain sectorial matrix functions were considered. In addition, attention was also paid to Toeplitz plus Hankel operators with piecewise almost periodic symbols and with scalar symbols having n points of standard almost periodic discontinuities. In the entire thesis a main goal is to obtain Fredholm properties description of those classes of operators. To deduce the lateral or both sided invertibility theory for Wiener-Hopf plus Hankel operators with AP matrix symbols was introduced the notion of an AP asymmetric factorization. In this framework were given sufficient conditions for the lateral and both sided invertibility of the Wiener-Hopf plus Hankel operators with matrix AP symbols. For such kind of operators were also exhibited generalized inverses for all the possible cases. For the Wiener-Hopf-Hankel operators with matrix SAP and PAP symbols the Fredholm property and a formula for the sum of the Fredholm indices of these Wiener-Hopf plus Hankel and Wiener-Hopf minus Hankel operators were derived. A stronger version of these results was obtained by using the generalized right AP factorization. It was analyzed the Wiener-Hopf-Hankel operators with symbols presenting some even properties, and also with Fourier symbols which contain sectorial matrices. In addition, for Wiener-Hopf-Hankel operators, it was obtained a corresponding result to the classical theorem by Douglas and Sarason known for Toeplitz operators with sectorial and unitary valued symbols. Necessary and sufficient condition for the Wiener-Hopf plus Hankel operators with L∞ symbols to be Fredholm (or invertible) were also derived. To obtain such a result we dealt with certain odd asymmetric factorization of the Fourier symbols. The Toeplitz plus Hankel operators generated by symbols which have n points of standard almost periodic discontinuities were also considered. Conditions were obtained under which these operators are right-invertible and with infinite kernel dimension, left-invertible and with infinite cokernel dimension or simply not normally solvable. We also focused our attention to Toeplitz plus Hankel operators with piecewise almost periodic matrix symbols. For such operators necessary and sufficient conditions were obtained to have the Fredholm property. This was done using a symbol calculus approach, certain auxiliary paired operators with semi-almost periodic symbols, and several equivalence after extension operator relations.

This thesis deals with generic determinacy and the number of equilibria for infinite dimensional economies. Our work could be seen as an infinite-dimensional analogue of Dierker and Dierker (1972) by characterising equilibria of an economy as a zero of the aggregate excess demand and studying its transversality. In this case, we can use extensions of the Sard-Smale theorem. Assuming separable utilities we give a new proof of generic determinacy of equilibria. We define regular price systems in this setting and show that an economy is regular if and only if its associated excess demand function only has regular equilibrium prices. We also define the infinite equilibrium manifold à la Balasko and show that it has the structure of a Banach manifold. We provide conditions that guarantee global uniqueness of equilibria for smooth infinite economies. We do this by introducing to the economic literature the notion of Z-Rothe vector fields that will allow us to construct an index theorem à la Dierker (1972); this shows that the number of equilibria is odd and in particular gives a new proof of existence. Extending the finite dimensional results of Balasko (1988), we characterise the equilibrium manifold as a covering space of the set of economies and we study global conditions under which the natural projection map is a diffeomorphism. We finally study the effects that critical equilibria have on the global invertibility of the natural projection map.

This text is concerned with the Fredholm theory and stable approximation of bounded linear operators generated by a class of infinite matrices $(a_{ij})$ that are either banded or have certain decay properties as one goes away from the main diagonal. The operators are studied on $\ell^p$ spaces of functions $\Z^N\to X$, where $p\in[1,\infty]$, $N\in\N$ and $X$ is a complex Banach space. The latter means that our matrix entries $a_{ij}$ are indexed by multiindices $i,j\in\Z^N$ and that every $a_{ij}$ is itself a bounded linear operator on $X$. Our main focus lies on the case $p=\infty$, where new results are derived, and it is demonstrated in both general theory and concrete operator equations from mathematical physics how advantage can be taken of these new $p=\infty$ results in the general case $p\in[1,\infty]$.

Le contexte général de cette thèse est celui de l'extension de la théorie des opérateurs elliptiques, bien connue dans le cadre lisse, à des domaines dits singuliers. Les méthodes utilisées reposent d'une part sur l'emploi d'algèbres d'opérateurs et d'outils issus de la géométrie non commutative, d'autre part sur l'introduction de calculs pseudodifférentiels adaptés à la géométrie du domaine, souvent via un groupoïde qui résout les singularités. La première partie de la thèse s'intéresse à l'étude d'une classe particulière de ces groupoïdes, dits Fredholm, qui donnent un cadre très favorable à l'analyse des opérateurs elliptiques. Un des résultats majeurs obtenu est que cette propriété de Fredholm est locale, au sens où elle ne dépend que des restrictions du groupoïde à un nombre suffisant d'ouverts. Dans le même esprit, nous considérons avec C. Carvalho et Y. Qiao des groupoïdes obtenus comme recollements d'actions de groupes, et étudions en particulier un groupoïde adapté à l'étude des opérateurs potentiels de couche. Je conclus cette partie avec la résolution d'un problème aux limites pour un domaine à singularité de type cusp rotationnel. La seconde partie s'intéresse aux opérateurs équivariants sur des variétés compactes, sous l'action d'un groupe fini. On répond à la question suivante : étant donnée une représentation irréductible du groupe, à quelle condition un opérateur différentiel est-il Fredholm entre les composantes isotypiques correspondantes des espaces de Sobolev ? Dans un travail commun avec A. Baldare, M. Lesch et V. Nistor, nous définissons une notion correspondante d'ellipticité associée à une représentation irréductible fixée et montrons qu'elle caractérise les opérateurs de Fredholm
This thesis is set in the general context of extending the theory of elliptic operators, well-understood in the smooth setting, to so-called singular domains. The methods used rely on operator algebras and tools coming from non commutative geometry, together with suitable pseudodifferential calculi that are often built from a groupoid adapted to the particular geometry of the problem. The first part of the thesis deals with the general investigation of a particular class of such groupoids, called Fredholm, that provide a very good setting for the study of elliptic operators. One of the major results proved here is that this Fredholm property is local, in the sense that it only depends on the restrictions of the groupoid to sufficiently many open subsets. In the same spirit, we study with C. Carvalho and Y. Qiao groupoids whose local structure is given by gluing group actions, and consider in particular a groupoid suited to the study of layer potential operators. This part concludes with a well-posedness result for a boundary value problem on a domain with a rotational cusp. The second part deals with equivariant operators on a compact manifold, acted upon by a finite group. We answer the following question: given an irreducible representation of the group, under which condition is a differential operator Fredholm between the corresponding isotypical components of the Sobolev spaces? In a joint work with A. Baldare, M. Lesch and V. Nistor, we introduce a corresponding notion of ellipticity associated with some fixed irreducible representation, and show that it characterizes Fredholm operators

Dans cette thèse on prouve des résultats analytiques sur la théorie cohomotopique de Seiberg-Witten pour des 4-variétes Riemanniennes Spinc(4) a bouts périodiques, (X,g,τ). Nos résultats montrent, que sur certaines conditions techniques en (X, g, τ ),, cette nouvelle version est cohérente et mène a des invariants de Seiberg-Witten.Premièrement, en utilisant le critère de Taubes pour des operateurs périodiques dans des variétes a bouts périodiques, on montre que pour une 4-varieté Riemmanienne a bouts périodiques (X, g) vérifiant certaines conditions topologiques, le Laplacian ∆+ : L2(Λ2+) → L2(Λ2+) est un opérateur de Fredholm. On prouve une décomposition de type Hodge pour des 1-formes de X, a poids positif.Ensuite on prouve, en assumant certaines conditions topologiques et courbure scalaire non-negative sur les bouts, que l'opérateur de Dirac associé a une connection périodique (ASD a l'infini) est Fredholm.Dans la deuxième partie de la thèse on démontre un isomorphisme entre le groupe de cohomologie de de Rham Hd1R(X,iR), et le groupe harmonique intervenant dans la decomposition de Hodge des 1-formes de X a poids positif. On prouve l'existence de deux séquences exactes courtes liant le groupe de jauge de l'espace de modules de Seiberg-Witten et le groupe de cohomologie H1(X, 2πiZ).Dans la troisième partie on prouve les principaux résultats: la coercitivité de l'application de Seiberg-Witten et la compacité de l'espace de moduli pour une 4-varieté a bouts périodiques (X, g, τ ), vérifiant les conditions mentionnées plus haut.Finalment, utilisant la coercivité, on montre l'existence d'un invariant cohomotopique de type Seiberg- Witten type associé a (X, g, τ )
In this thesis we prove analytic results about a cohomotopical Seiberg-Witten theory for a Riemannian, Spinc(4) 4-manifold with periodic ends, (X,g,τ) . Our results show that, under certain technical assumptions on (X, g, τ ), this new version is coher- ent and leads to Seiberg-Witten type invariants for this new class of 4-manifolds.First, using Taubes criteria for end-periodic operators on manifolds with periodic ends, we show that, for a Riemannian 4-manifold with periodic ends (X, g), verifying certain topological conditions, the Laplacian ∆+ : L2(Λ2+) → L2(Λ2+) is a Fredholm operator. This allows us to prove an important Hodge type decomposition for positively weighted Sobolev 1-forms on X.We prove, assuming non-negative scalar curvature on each end and certain technical topological conditions, that the associated Dirac operator associated with an end-periodic connection (which is ASD at infinity) is Fredholm.In the second part of the thesis we establish an isomorphism between be- tween the de Rham cohomology group, Hd1R(X,iR) (which is a topological in- variant of X) and the harmonic group intervening in the above Hodge type decomposition of the space of positively weighted 1-forms on X. We also prove two short exact sequences relating the gauge group of our Seiberg-Witten moduli problem and the cohomology group H1(X, 2πiZ).In the third part, we prove our main results: the coercivity of the Seiberg-Witten map and compactness of the moduli space for a 4-manifold with periodic ends (X,g,τ) verifying the above conditions.Finally, using our coercitivity property, we show that a Seiberg-Witten type cohomotopy invariant associated to (X, g, τ ) can be defined

Die vorliegende Arbeit ist der Untersuchung von Operatorfolgen gewidmet, die typischerweise bei der Anwendung von Approximationsverfahren auf stetige lineare Operatoren entstehen. Dabei stehen die Stabilität der Folgen sowie das asymptotische Verhalten gewisser Charakteristika wie Normen, Konditionszahlen, Fredholmeigenschaften und Pseudospektren im Mittelpunkt. Das Hauptaugenmerk liegt auf der Entwicklung der Theorie für Operatoren auf Banachräumen. Hierbei bildet ein dafür geeigneter Konvergenzbegriff, die sogenannte P-starke Konvergenz, den Ausgangspunkt, welcher das Studium der gewünschten Eigenschaften in einer erstaunlichen Allgemeinheit gestattet. Die erzielten Resultate kommen, neben einer Reihe weiterer Anwendungen, insbesondere für das Projektionsverfahren für banddominierte Operatoren zum Einsatz.

O objetivo principal desta dissertação é apresentar o fluxo espectral de um caminho de operadores de Fredholm auto-adjuntos em um espaço de Hilbert e suas propriedades. Pelos resultados clássicos de teoria espectral, sabemos que se H é um espaço de Hilbert e L : H → H é um operador linear, limitado e auto-adjunto, H pode ser escrito como soma direta ortogonal H+(L)⊕ H-(L)⊕ Ker L, onde H+(L) e H-(L) são os subespaços espectrais positivo e negativo de L, respectivamente. No trabalho damos uma definição de fluxo espectral baseada na decomposição acima, aprofundando as conexões deste conceito com a teoria espectral dos operadores de Fredholm em espaços de Hilbert. Entre as propriedades do fluxo espectral, será analisada a invariância homotópica que se apresenta em várias formas. Veremos o conceito de índice de Morse relativo, que estende o clássico índice de Morse, e sua relação com o fluxo espectral. A construção do fluxo espectral dada neste trabalho segue a abordagem de P. M. Fitzpatrick, J. Pejsachowicz e L. Recht em [9].
The main purpose of this dissertation is to present the spectral flow of a path of selfadjoint Fredholm operators in a Hilbert space and its properties. By classical results in spectral theory, we know that, if H is a Hilbert space and L : H → H is a bounded self-adjoint linear operator, H may be written as the following orthogonal direct sum H = H+(L)⊕ H-(L)⊕ Ker L, where H+(L) and H-(L) are the positive and negative spectral subspaces of L, respectively. In this work we give a definition of spectral flow which is based on the above splitting, examining in depth the connection between this concept and the spectral theory of Fredholm operators in Hilbert spaces. Among the properties of the spectral flow we will analyze the homotopic invariance, which appears on different ways. We will see the concept of relative Morse index, which generalize the classical Morse index, and its relation with the spectral flow. The construction of the spectral flow given in this work follows the approach of P. M. Fitzpatrick, J. Pejsachowicz and L. Recht in [9].

Estudamos a álgebra de comparação do círculo, descrevemos o espaço de Gelfand do seu quociente pelos compactos e damos uma fórmula para o índice dos seus operadores de Fredholm. Depois generalizamos o resultado para as matrizes com elementos na álgebra de comparação e damos uma aplicação para operadores diferenciais no círculo.
We study the comparison algebra on the circle, we describe the Gelfand space of its quotient by the compacts and we give a formula to compute the index of its Fredholm operators. After that, we generalize the result to the matrices with entries in the comparison algebra and give an application to differential opperators in the circle.

We construct elliptic theory in the subspaces, determined by pseudodifferential projections. The finiteness theorem as well as index formula are obtained for elliptic operators acting in the subspaces. Topological (K-theoretic) aspects of the theory are studied in detail.

In the paper we study the possibility to represent the index formula for spectral boundary value problems as a sum of two terms, the first one being homotopy invariant of the principal symbol, while the second depends on the conormal symbol of the problem only. The answer is given in analytical, as well as in topological terms.

Nous nous intéressons à des équations différentielles stochastiques obtenues en perturbant par un bruit blanc des équations différentielles ordinaires admettant N orbites périodiques asymptotiquement stables. Nous construisons une chaîne de Markov à temps discret et espace d’états continu appelée application de Poincaré aléatoire qui hérite du comportement métastable du système. Nous montrons que ce processus admet exactement N valeurs propres qui sont exponentiellement proches de 1 et nous donnons des expressions pour ces valeurs propres et les fonctions propres associées en termes de fonctions committeurs dans les voisinages des orbites périodiques. Nous montrons également que ces valeurs propres sont bien séparées du reste du spectre. Chacune de ces valeurs propres exponentiellement proche de 1 est également reliée à un temps d’atteinte de ces voisinages. De plus, les N valeurs propres exponentiellement proches de 1 et fonctions propres à gauche et à droite associées peuvent être respectivement approchées par des valeurs propres principales, des distributions quasi-stationnaires, et des fonctions propres principales à droite de processus tués quand ils atteignent ces voisinages. Les preuves reposent sur une représentation de type Feynman–Kac pour les fonctions propres, la transformée harmonique de Doob, la théorie spectrale des opérateurs compacts et une propriété de type équilibré détaillé satisfaite par les fonctions committeurs
We consider stochastic differential equations, obtained by adding weak Gaussian white noise to ordinary differential equations admitting N asymptotically stable periodic orbits. We construct a discrete-time,continuous-space Markov chain, called a random Poincaré map, which encodes the metastable behaviour of the system. We show that this process admits exactly N eigenvalues which are exponentially close to 1,and provide expressions for these eigenvalues and their left and right eigenfunctions in terms of committorfunctions of neighbourhoods of periodic orbits. We also provide a bound for the remaining part of the spectrum. The eigenvalues that are exponentially close to 1 and the right and left eigenfunctions are well-approximated by principal eigenvalues, quasistationary distributions, and principal right eigenfunctions of processes killed upon hitting some of these neighbourhoods. Each eigenvalue that is exponentially close to 1is also related to the mean exit time from some metastable neighborhood of the periodic orbits. The proofsrely on Feynman–Kac-type representation formulas for eigenfunctions, Doob’s h-transform, spectral theory of compact operators, and a recently discovered detailed balance property satisfied by committor functions

In dieser Arbeit wird das asymptotische Verhalten der Approximationszahlen für Operatorfolgen aus einer speziellen Klasse von Banachalgebren untersucht. Es werden bemerkenswerte Eigenschaften der Folgen und der Approximationszahlen ihrer Operatoren gezeigt, darunter die so genannte splitting-Eigenschaft. Ein typisches Beispiel solcher Operatorfolgen stellen die Finite Sections von Toeplitzoperatoren dar, die exemplarisch behandelt werden. Dabei werden hier auch die Folgenräume l1 und l-unendlich betrachtet.

Die abgeschlossenen Erweiterungen der sogenannten geometrischen Operatoren (Spin-Dirac, Gauß-Bonnet und Signatur-Operator) auf Mannigfaltigkeiten mit metrischen Hörnern sind Fredholm-Operatoren und ihr Index wurde von Matthias Lesch, Norbert Peyerimhoff und Jochen Brüning berechnet. Es wurde gezeigt, dass die Einschränkungen dieser drei Operatoren auf eine punktierte Umgebung des singulären Punkts unitär äquivalent zu irregulär singulären Operator-wertigen Differentialoperatoren erster Ordnung sind. Die Lösungsoperatoren der dazugehörigen Differentialgleichungen definierten eine Parametrix, mit deren Hilfe die Fredholmeigenschaft bewiesen wurde. In der vorliegenden Doktorarbeit wird eine Klasse von irregulären singulären Differentialoperatoren erster Ordnung, genannt Horn-Operatoren, eingeführt, die die obigen Beispiele verallgemeinern. Es wird bewiesen, dass ein elliptischer Differentialoperator erster Ordnung, dessen Einschränkung auf eine punktierte Umgebung des singulären Punkts unitär äquivalent zu einem Horn-Operator ist, Fredholm ist, und sein Index wird berechnet. Schließlich wird dieser abstrakte Index-Satz auf geometrische Operatoren auf Mannigfaltigkeiten mit "multiply warped product"-Singularitäten angewendet, welche eine wesentliche Verallgemeinerung der metrischen Hörner darstellen.
The closed extensions of geometric operators (Spin-Dirac, Gauss-Bonnet and Signature operator) on a manifold with metric horns are Fredholm operators, and their indices were computed by Matthias Lesch, Norbert Peyerimhoff and Jochen Brüning. It was shown that the restrictions of all three operators to a punctured neighbourhood of the singular point are unitary equivalent to a class of irregular singular operator-valued differential operators of first order. The solution operators of the corresponding differential equations defined a parametrix which was applied to prove the Fredholm property. In this thesis a class of irregular singular differential operators of first order - called horn operators - is introduced that extends the examples mentioned above. It is proved that an elliptic differential operator of first order whose restriction to the neighbourhood of the singular point is unitary equivalent to a horn operator is Fredholm and its index is computed. Finally, this abstract index theorem is applied to compute the indices of geometric operators on manifolds with multiply warped product singularities that extend the notion of metric horns considerably.

The aim of this thesis is to give a mathematical framework for scattering of electromagnetic waves by rough surfaces. We prove that the Maxwell transmission problem with a weakly Lipschitz interface,in finite energy norms, is well posed in Fredholm sense for real frequencies. Furthermore, we give precise conditions on the material constants ε, μ and σ and the frequency ω when this transmission problem is well posed. To solve the Maxwell transmission problem, we embed Maxwell’s equations in an elliptic Dirac equation. We develop a new boundary integral method to solve the Dirac transmission problem. This method uses a boundary integral operator, the rotation operator, which factorises the double layer potential operator. We prove spectral estimates for this rotation operator in finite energy norms using Hodge decompositions on weakly Lipschitz domains. To ensure that solutions to the Dirac transmission problem indeed solve Maxwell’s equations, we introduce an exterior/interior derivative operator acting in the trace space. By showing that this operator commutes with the two basic reflection operators, we are able to prove that the Maxwell transmission problem is well posed. We also prove well-posedness for a class of oblique Dirac transmission problems with a strongly Lipschitz interface, in the L_2 space on the interface. This is shown by employing the Rellich technique, which gives angular spectral estimates on the rotation operator.

La méthode d'intégration produit a été proposée pour résoudre des équations linéaires de Fredholm de deuxième espèce singulières dont la solution exacte est régulière, au moins continue. Dans ce travail on adapte cette méthode à des équations dont la solution est juste intégrable. On étudie également son extension au cas non linéaire posé dans l'espace des fonctions intégrables. Ensuite, on propose une autre manière de mettre en oeuvre la méthode d'intégration produit : on commence par linéariser l'équation par une méthode de type Newton puis on discrétise les itérations de Newton par la méthode d'intégration produit
The product integration method has been proposed for solving singular linear Fredholm equations of the second kind whose exact solution is smooth, at least continuous. In this work, we adapt this method to the case where the solution is only integrable. We also study the nonlinear case in the space of integrable functions. Then, we propose a new version of the method in the nonlinear framework : we first linearize the eqaution by a Newton type method and then discretize the Newton iterations by the product integration method

La détermination de la forme d'un obstacle élastique immergé dans un milieu fluide à partir de mesures du champ d'onde diffracté est un problème d'un vif intérêt dans de nombreux domaines tels que le sonar, l'exploration géophysique et l'imagerie médicale. A cause de son caractère non-linéaire et mal posé, ce problème inverse de l'obstacle (IOP) est très difficile à résoudre, particulièrement d'un point de vue numérique. De plus, son étude requiert la compréhension de la théorie du problème de diffraction direct (DP) associé, et la maîtrise des méthodes de résolution correspondantes. Le travail accompli ici se rapporte à l'analyse mathématique et numérique du DP élasto-acoustique et de l'IOP. En particulier, nous avons développé un code de simulation numérique performant pour la propagation des ondes associée à ce type de milieux, basé sur une méthode de type DG qui emploie des éléments finis d'ordre supérieur et des éléments courbes à l'interface afin de mieux représenter l'interaction fluide-structure, et nous l'appliquons à la reconstruction d'objets par la mise en oeuvre d'une méthode de Newton régularisée.

In the classical theory of operators on a Banach space a beautiful interplay exists between Riesz and Fredholm theory, and the theory of traces and de¬terminants for operator ideals. In this thesis we obtain a complete Riesz de¬composition theorem for Riesz elements in a semi prime Banach algebra and on the other hand extend the existing theory of traces and determinants to a more general setting of Banach algebras. In order to obtain some of these results we use the notion of finite multiplicity of spectral points to give a characterization of the essential spec¬trum for elements in a Banach algebra. As an immediate corollary we obtain the well-known characterization of Riesz elements namely that their non-zero spectral points are isolated and of finite multiplicities. In the final chapter of the thesis we use Plemelj's type formulas to define a determinant on the ideal of finite rank elements and show that it extends continuously to the ideal of nuclear elements.
Thesis (PhD (Mathematics))--University of Pretoria, 2006.
Mathematics and Applied Mathematics
unrestricted

Die vorliegende Arbeit ist der Untersuchung von Operatorfolgen gewidmet, die typischerweise bei der Anwendung von Approximationsverfahren auf stetige lineare Operatoren entstehen. Dabei stehen die Stabilität der Folgen sowie das asymptotische Verhalten gewisser Charakteristika wie Normen, Konditionszahlen, Fredholmeigenschaften und Pseudospektren im Mittelpunkt. Das Hauptaugenmerk liegt auf der Entwicklung der Theorie für Operatoren auf Banachräumen. Hierbei bildet ein dafür geeigneter Konvergenzbegriff, die sogenannte P-starke Konvergenz, den Ausgangspunkt, welcher das Studium der gewünschten Eigenschaften in einer erstaunlichen Allgemeinheit gestattet. Die erzielten Resultate kommen, neben einer Reihe weiterer Anwendungen, insbesondere für das Projektionsverfahren für banddominierte Operatoren zum Einsatz.

Indiana University-Purdue University Indianapolis (IUPUI)
We study the one-parameter family of determinants $det(I-\gamma K_{PII}),\gamma\in\mathbb{R}$ of an integrable Fredholm operator $K_{PII}$ acting on the interval $(-s,s)$ whose kernel is constructed out of the $\Psi$-function associated with the Hastings-McLeod solution of the second Painlev\'e equation. In case $\gamma=1$, this Fredholm determinant describes the critical behavior of the eigenvalue gap probabilities of a random Hermitian matrix chosen from the Unitary Ensemble in the bulk double scaling limit near a quadratic zero of the limiting mean eigenvalue density. Using the Riemann-Hilbert method, we evaluate the large $s$-asymptotics of $\det(I-\gamma K_)$ for all values of the real parameter $\gamma$.

The aim of this thesis is to give a mathematical framework for scattering of electromagnetic waves by rough surfaces. We prove that the Maxwell transmission problem with a weakly Lipschitz interface, in finite energy norms, is well posed in Fredholm sense for real frequencies. Furthermore, we give precise conditions on the material constants ε, μ and σ and the frequency ω when this transmission problem is well posed. To solve the Maxwell transmission problem, we embed Maxwell’s equations in an elliptic Dirac equation. We develop a new boundary integral method to solve the Dirac transmission problem. This method uses a boundary integral operator, the rotation operator, which factorises the double layer potential operator. We prove spectral estimates for this rotation operator in finite energy norms using Hodge decompositions on weakly Lipschitz domains. To ensure that solutions to the Dirac transmission problem indeed solve Maxwell’s equations, we introduce an exterior/interior derivative operator acting in the trace space. By showing that this operator commutes with the two basic reflection operators, we are able to prove that the Maxwell transmission problem is well posed. We also prove well-posedness for a class of oblique Dirac transmission problems with a strongly Lipschitz interface, in the L_2 space on the interface. This is shown by employing the Rellich technique, which gives angular spectral estimates on the rotation operator.

In dieser Arbeit wird das asymptotische Verhalten der Approximationszahlen für Operatorfolgen aus einer speziellen Klasse von Banachalgebren untersucht. Es werden bemerkenswerte Eigenschaften der Folgen und der Approximationszahlen ihrer Operatoren gezeigt, darunter die so genannte splitting-Eigenschaft. Ein typisches Beispiel solcher Operatorfolgen stellen die Finite Sections von Toeplitzoperatoren dar, die exemplarisch behandelt werden. Dabei werden hier auch die Folgenräume l1 und l-unendlich betrachtet.