Cette thèse est consacrée à la théorie spectrale de quelques opérateurs de transport et le comportement asymptotique (pour les temps grands) des solutions des problèmes de Cauchy gouvernés par ces derniers. Dans la première partie, on s'est intéressé aux propriétés spectrales des opérateurs d'advection et de transport des neutrons dans le cadre multidimensionnel pour des conditions aux limites générales. Après avoir établi un résultat de compacité de type lemmes de moyenne indispensable dans notre analyse, on a donné entre autre une description fine du spectre asymptotique de l'opérateur de transport. Ce travail a été complété par l'étude des propriétés de régularité et le comportement asymptotique de la solution du problème de Cauchy gouverné par l'opérateur de transport étudié précédemment pour des conditions aux limites de type bounce-back plus un opérateur compact dans l'espace L^1. Ensuite, on a étudié le caractère bien posé et le comportement asymptotique de la solution d'une équation de transport des neutrons avec des sections efficaces non bornées. Contrairement à la première partie, l'analyse de ce problème nécessite l'usage d'une théorie de perturbation de Miyadera-Voigt pour les opérateurs non bornés. La dernière partie de ce travail porte sur un problème linéaire issu d'un modèle introduit en 1974 par Lebowitz et Rubinow décrivant la prolifération d'une population de cellules structuré par l'âge et la longueur du cycle. Notre analyse a porté sur le cas où la longueur du cycle maximale est infinie. / This thesis is devoted to the spectral theory and the time asymptotic behavior of the solution to Cauchy problems governed by various transport operators. In the first part, we discussed the spectral properties of streaming and transport operators in finite bodies with general boundary conditions. After establishing a compactness result essential to our analysis, we gave a fine description of the asymptotic spectrum of the transport operator. We also derive the regularity and the asymptotic behavior of the solution to Cauchy problem governed by the transport operator supplemented by bounce-back boundary conditions plus a compact operator in the space L^1. In the second part, we discussed the well-posedness and the asymptotic behavior of the solution to Cauchy problem governed by a singular transport operator. Unlike the first part, the analysis of this problem requires the use of Miyadera-Voigt perturbation theory for unbounded operators. In the last part of this work, a Cauchy problem governed by a linear operator introduced by Lebowitz and Rubinow describing a proliferating cell population structured by age and the cycle length was considered. Here our analysis was devoted to the case where the maximum cycle length is infinite.

Analysis of vibration of systems with uncertainty in material properties under the influence of a random forcing function is an active area of research. Especially the characterization based on mode shapes and frequencies of linear vibrating systems leads to much discussed random eigenvalue problem, which repeatedly appears while analyzing a number of engineering systems. Such analyses with conventional schemes for significant variation of system parameters for large systems are often not viable because of the high computational costs involved. Appropriate tools to reduce the size of stochastic vibrating systems and efficient response calculation are yet to mature. Among the mathematical tools used in this case, polynomial chaos formulation of uncertainties shows promise. But this comes with the implementation issue of solving large systems of nonlinear equations arising from Bubnov-Galerking projection in the formulation. This dissertation reports the study of such dynamic systems with uncertainties characterized by the probability distribution of eigen solutions under a stochastic finite element framework. In the context of structural vibration, the determination of appropriate modes to be considered in a stochastic framework is not straightforward. In this dissertation, at first the choice of dominant modes in stochastic framework is studied for vibration problems. A relative measure, based on the average energy contribution of each mode to the system, is developed. Further the interdependence of modes and the effect of the shape of the load on the choice of dominant modes are studied. Using these considerations, a hybrid algorithm is developed based on polynomial chaos framework for the response analysis of a structure with random mass and sickness and under the influence of random force. This is done by using modal truncation for response analysis with in a Monte Carlo loop. The algorithm is observed to be more efficient and achieves a high degree of accuracy compared to conventional techniques. Considering the fact that the Monte Carlo loops within the above mentioned hybrid algorithm is easily parallelizable, the efficient implementation of it depends on the SFE solution. The set of nonlinear equations arising from polynomial chaos formulation is solved using matrix-free Newton’s iteration using GMRES as linear solver. Solution of a large system using a iterative method like GMRES necessitates the use of a good preconditioner. Keeping focus on the par-allelizability of the algorithm, a number of efficient but cheap-to-construct preconditioners are developed and the most effective among them is chosen. The solution process is parallelized for large systems. The scalability of solution process in conjunction with the preconditioner is studied in details.

Rotating and non-rotating beams are widely used to model important engineering struc-tures. Hence, the vibration analyses of these beams are an important problem from a structural dynamics point of view. Depending on the beam dimensions, they are mod-eled using different beam theories. In most cases, the governing differential equations of these types of beams do not yield any simple closed-form solutions; hence we look for the inverse problem approach in determining the beam property variations given certain solutions. The long and slender beams are generally modeled using the Euler-Bernoulli beam theory. Under the premise of this theory, we study (i) the second mode tailoring of non-rotating beams having six different boundary conditions, (ii) closed-form solutions for free vibration analysis of free-free beams, (iii) closed-form solutions for free vibration analysis for gravity-loaded cantilever beams, (iv) closed-form solutions for free vibration analysis of rotating cantilever and pinned-free beams and (v) beams with shared eigen-pair. Short and thick beams are generally modeled using the Timoshenko beam theory. Here, we provide analytical closed-form solutions for the free vibration analysis of ro-tating non-homogeneous Timoshenko beams. The Rayleigh beam provides a marginal improvement over the Euler-Bernoulli beam theory without venturing into the math-ematical complexities of the Timoshenko beam theory. Under this theory, we provide closed-form solutions for the free vibration analysis of cantilever Rayleigh beams under three different axial loading conditions - uniform loading, gravity-loading and centrifu-gally loaded. We assume simple polynomial mode shapes which satisfy the different boundary conditions of a particular beam, and derive the corresponding beam property variations. In case of the shared eigenpair, we use the mode shape of a uniform beam which has a closed-form solution and use it to derive the stiffness distribution of a corresponding axially loaded beam having same length, mass variation and boundary condition. For the Timoshenko beam, we assume polynomial functions for the bending displacement and the rotation due to bending. The derived properties are demonstrated as benchmark analytical solutions for approximate and numerical methods used for the free vibration analysis of beams. They can also aid in designing actual beams for a pre-specified frequency or nodal locations in some cases. The effect of different parameters in the derived property variations and the bounds on the pre-specified frequencies and nodal locations are also studied for certain cases. The derived analytical solutions can also serve as a benchmark solution for different statistical simulation tools to find the probabilistic nature of the derived stiffness distri-bution for known probability distributions of the pre-specified frequencies. In presence of uncertainty, this flexural stiffness is treated as a spatial random field. For known probability distributions of the natural frequencies, the corresponding distribution of this field is determined analytically for the rotating cantilever Euler-Bernoulli beams. The derived analytical solutions are also used to derive the coefficient of variation of the stiffness distribution, which is further used to optimize the beam profile to maximize the allowable tolerances during manufacturing.

The solution of large and sparse eigenvalue problems arising from the discretization of the diffusion equation is considered. The multigroup<p>diffusion equation is discretized by means of the Nodal expansion Method (NEM) [9, 10]. A new formulation of the higher order NEM variants revealing the true nature of the problem, that is, a generalized eigenvalue problem, is proposed. These generalized eigenvalue problems are solved using the Jacobi-Davidson (JD) method<p>[26]. The most expensive part of the method consists of solving a linear system referred to as correction equation. It is solved using Krylov subspace methods in combination with aggregation-based Algebraic Multigrid (AMG) techniques. In that context, a particular<p>aggregation technique used in combination with classical smoothers, referred to as oblique geometric coarsening, has been derived. Its particularity is that it aggregates unknowns that<p>are not coupled, which has never been done to our<p>knowledge. A modular code, combining JD with an AMG preconditioner, has been developed. The code comes with many options, that have been tested. In particular, the instability of the Rayleigh-Ritz [33] acceleration procedure in the non-symmetric case has been underlined. Our code has also been compared to an industrial code extracted from ARTEMIS. / Doctorat en Sciences de l'ingénieur / info:eu-repo/semantics/nonPublished

Cette thèse est consacrée à l'étude d'équations de réaction-diffusion dans un environnement périodique en temps. Ces équations modélisent l'évolution d'une tumeur cancéreuse en présence d'un traitement qui correspond à une immunothérapie dans la première partie du manuscrit, et à une chimiothérapie cytotoxique dans la suite.On considère dans un premier temps des nonlinéarités périodiques en temps pour lesquelles 0 et 1 sont des états d'équilibre linéairement stables. On étudie l'unicité, la monotonie et la stabilité de fronts pulsatoires. On exhibe également des cas d'existence et de non-existence de telles solutions. Dans la deuxième partie de la thèse, on commence par travailler sur des nonlinéarités périodiques en temps qui sont la somme d'une fonction positive traduisant la croissance de la tumeur et d'un terme de mort de cellules cancéreuses du au traitement. On s'intéresse aux états d'équilibres de telles nonlinéarités, et on va déduire de cette étude des propriétés de propagation de perturbations et l'existence de fronts pulsatoires. On raffine ensuite le modèle en considérant des nonlinéarités qui sont la somme d'une fonction asymptotiquement périodique en temps et d'un terme perturbatif. On prouve notamment que les propriétés relatives à la propagation de perturbations restent valables dans ce cadre là. Pour finir, on s'intéresse à l'influence du protocole de traitement. / This phD thesis investigates reaction-diffusion equations in a time periodic environment. These equations model the evolution of a cancerous tumor in the presence of a treatment that corresponds to an immunotherapy in the firs part of the manuscript, and to a cytotoxic chemotherapy after. We begin by considering time-periodic nonlinearities for which 0 and 1 are linearly stable equilibrium states. We study uniqueness, monotonicity and stability of pulsating fronts. We also provide some conditions for the existence and non-existence of such solutions.In the second part of the manuscript, we begin by working on time-periodic nonlinearities which are the sum of a positive function which stands for the growth of the tumor in the absence of treatment and of a death term of cancerous cells due to treatment. We are interested in equilibrium states of such nonlinearities, and we will infer from this study spreading properties and existence of pulsating fronts. We then refine the model by considering nonlinearities which are the sum of an asymptotic periodic nonlinearity and of a small perturbation. In particular we prove that the spreading properties remain valid in this case. To finish, we are interested in the influence of the protocol of the treatment.

In this thesis, a new method is followed to design an active control system which imparts viscoelastic phenomenological damping in an elastic structure. Properties of a hypothetical viscoelastic system are used to design an active feedback controller for an undamped structural system with distributed sensor, actuator and controller. The variational structure is projected on a solution space of a closed-loop system involving a weakly damped structure with distributed sensor and actuator with controller. These controller components assign the phenomenology based on internal strain rate damping parameter of a viscoelastic system to the undamped elastic structure. An elastic cantilever beam with proportional-derivative controller and displacement feedback is considered in all the design formulations. In the first part of the research, a closed-loop control system is designed using two time domain modern control system design methods, pole placement and optimal pole placement, which use viscoelastic damping parameter. Equation of motion of a viscoelastic system is employed to synthesize the desired closed-loop poles. Desired poles are then assigned to an elastic beam with an active controller. Time domain finite element formulation is used without considering actuator-sensor dynamics and the effect of transducer locations. Characteristics of closed-loop system gains are found as a function of desired damping parameter and realization of damping have been analyzed with closed loop system pole positions. The next part consists of a novel frequency domain active control system design to impart desired viscoelastic characteristics, which uses spectral method and the exact dynamic stiffness matrix of the system. In the first case, a sub-optimal local control system for a cantilever beam, with collocated actuator and sensor is designed. In the second case, a closed-loop local controller for an elastic system with non-collocated transducers is designed. Next, a global controller for non-collocated arrangement of sensor-actuator is designed by considering all the degrees-of freedom in the system, which leads to solving an eigenvalue problem. The reason for the failure of the collocated arrangement in global control is also explained. In this novel control system design method transducer dynamics and locations are considered in the formulation. In frequency domain design, the frequency responses of the system show satisfactory performance of the closed-loop elastic system. The closed-loop system is able to reproduce the desired viscoelastic characteristics as targeted in the design. Optimal and sub-optimal system gains are found as functions of transducer locations, transducer properties, excitation frequency and internal strain rate damping parameter of a hypothetical viscoelastic system. Performance of the closed loop system is established by comparing the specific damping capacity of the hypothetical viscoelastic system with that of the closed-loop elastic system. The novel frequency domain method is simple, accurate, efficient and can be extended to complex structures to achieve desired damping. The method can be a better way of designing structures with variable stiffness which has research potential in designing morphing airplanes/spacecrafts. The ultimate goal of this research is that, if this design method is applied to practical applications such as aircraft wings, where vibration is undesirable, one would be able to achieve strength and desired damping characters simultaneously.

Cette thèse concerne les phénomènes de propagation de certaines équations d'évolution dans des habitats périodiques. Dans la première partie, nous étudions les phénomènes d'expansion de certaines équations d'intégro-différence spatialement périodiques. Tout d'abord, nous établissons une théorie générale sur l'existence des vitesses de propagation pour des systèmes d'évolution noncompacts, sous l'hypothèse que les systèmes linéarisés ont des valeurs propres principales. Ensuite, nous introduisons la notion d'irréductibilité uniforme des mesures de Radon finies sur le cercle. On démontre que tout opérateur de convolution généré par une telle mesure admet une valeur propre principale. Enfin, nous prouvons l'existence de vitesses de propagation pour certains équations d'intégro-différence avec des noyaux de dispersion uniformément irréductibles. Dans la deuxième partie, nous étudions les phénomènes de propagation de front pour des équations de réaction-diffusion spatialement périodiques avec des non-linéarités bistables. Nous nous concentrons d'abord sur les solutions de type fronts pulsatoires. Sous diverses hypothèses, il est prouvé que les fronts pulsatoires existent lorsque la période spatiale est petite ou grande. Nous caractérisons aussi le signe des vitesses et nous montrons la stabilité exponentielle globale des fronts pulsatoires de vitesse non nulle. Nous étudions ensuite les solutions de type fronts de transition. Sous des hypothèses convenables, on prouve que les fronts de transition se ramènent aux fronts pulsatoires avec une vitesse non nulle. Mais nous montrons aussi l'existence de nouveaux types de fronts de transition qui ne sont pas des fronts pulsatoires. / This dissertation is concerned with propagation phenomena of some evolution equations in periodic habitats. The main results consist of the following two parts. In the first part, we investigate the spatial spreading phenomena of some spatially periodic integro-difference equations. Firstly, we establish a general theory on the existence of spreading speeds for noncompact evolution systems, under the hypothesis that the linearized systems have principal eigenvalues. Secondly, we introduce the notion of uniform irreducibility for finite Radon measures on the circle. It is shown that, any generalized convolution operator generated by such a measure admits a principal eigenvalue. Finally, applying the above general theories, we prove the existence of spreading speeds for some integro-difference equations with uniformly irreducible dispersal kernels. In the second part, we study the front propagation phenomena of spatially periodic reaction-diffusion equations with bistable nonlinearities. Firstly, we focus on the propagation solutions in the class of pulsating fronts. It is proved that, under various assumptions on the reaction terms, pulsating fronts exist when the spatial period is small or large. We also characterize the sign of the front speeds and we show the global exponential stability of the pulsating fronts with nonzero speed. Secondly, we investigate the propagation solutions in the larger class of transition fronts. It is shown that, under suitable assumptions, transition fronts are reduced to pulsating fronts with nonzero speed. But we also prove the existence of new types of transition fronts which are not pulsating fronts.

In Markov chain models in finance and healthcare a transition matrix over a certain time interval is needed but only a transition matrix over a longer time interval may be available. The problem arises of determining a stochastic $p$th root of astochastic matrix (the given transition matrix). By exploiting the theory of functions of matrices, we develop results on the existence and characterization of stochastic $p$th roots. Our contributions include characterization of when a real matrix hasa real $p$th root, a classification of $p$th roots of a possibly singular matrix,a sufficient condition for a $p$th root of a stochastic matrix to have unit row sums,and the identification of two classes of stochastic matrices that have stochastic $p$th roots for all $p$. We also delineate a wide variety of possible configurationsas regards existence, nature (primary or nonprimary), and number of stochastic roots,and develop a necessary condition for existence of a stochastic root in terms of the spectrum of the given matrix. On the computational side, we emphasize finding an approximate stochastic root: perturb the principal root $A^{1/p}$ or the principal logarithm $\log(A)$ to the nearest stochastic matrix or the nearest intensity matrix, respectively, if they are not valid ones;minimize the residual $\normF{X^p-A}$ over all stochastic matrices $X$ and also over stochastic matrices that are primary functions of $A$. For the first two nearness problems, the global minimizers are found in the Frobenius norm. For the last two nonlinear programming problems, we derive explicit formulae for the gradient and Hessian of the objective function $\normF{X^p-A}^2$ and investigate Newton's method, a spectral projected gradient method (SPGM) and the sequential quadratic programming method to solve the problem as well as various matrices to start the iteration. Numerical experiments show that SPGM starting with the perturbed $A^{1/p}$to minimize $\normF{X^p-A}$ over all stochastic matrices is method of choice.Finally, a new algorithm is developed for computing arbitrary real powers $A^\a$ of a matrix $A\in\mathbb{C}^{n\times n}$. The algorithm starts with a Schur decomposition,takes $k$ square roots of the triangular factor $T$, evaluates an $[m/m]$ Pad\'e approximant of $(1-x)^\a$ at $I - T^$, and squares the result $k$ times. The parameters $k$ and $m$ are chosen to minimize the cost subject to achieving double precision accuracy in the evaluation of the Pad\'e approximant, making use of a result that bounds the error in the matrix Pad\'e approximant by the error in the scalar Pad\'e approximant with argument the norm of the matrix. The Pad\'e approximant is evaluated from the continued fraction representation in bottom-up fashion, which is shown to be numerically stable. In the squaring phase the diagonal and first superdiagonal are computed from explicit formulae for $T^$, yielding increased accuracy. Since the basic algorithm is designed for $\a\in(-1,1)$, a criterion for reducing an arbitrary real $\a$ to this range is developed, making use of bounds for the condition number of the $A^\a$ problem. How best to compute $A^k$ for a negative integer $k$ is also investigated. In numerical experiments the new algorithm is found to be superior in accuracy and stability to several alternatives,including the use of an eigendecomposition, a method based on the Schur--Parlett\alg\ with our new algorithm applied to the diagonal blocks and approaches based on the formula $A^\a = \exp(\a\log(A))$.

Cette thèse est consacrée à l'étude du spectre de l'opérateur de Laplace avec conditions de Dirichlet dans différents domaines du plan ou de l'espace. Dans un premier temps on s'intéresse à des triangles asymptotiquement plats et des cônes de petite ouverture. Ces problèmes admettent une reformulation semi-classique et nous donnons des développements asymptotiques à tout ordre des premières valeurs et fonctions propres. Ce type de résultat est déjà connu pour des domaines minces à profil régulier. Pour les triangles et les cônes, on prouve que le problème admet maintenant deux échelles. Dans un second temps, on étudie une famille de couches coniques indexées par leur ouverture. Là encore, on s'intéresse à la limite semi-classique quand l'ouverture tend vers zéro: on donne un développement asymptotique à deux termes des premières valeurs propres et on démontre un résultat de localisation des fonctions propres associées. Nous donnons également, à ouverture fixée, un équivalent du nombre de valeurs propres sous le seuil du spectre essentiel. / This thesis deals with the spectrum of the Dirichlet Laplacian in various two or three dimensional domains. First, we consider asymptotically flat triangles and cones with small aperture. These problems admit a semi-classical formulation and we provide asymptotic expansions at any order for the first eigenvalues and the associated eigenfunctions. These type of results is already known for thin domains with smooth profiles. For triangles and cones, we show that the problem admits now two different scales. Second, we study a family of conical layers parametrized by their aperture. Again, we consider the semi-classical limit when the aperture tends to zero: We provide a two-term asymptotics of the first eigenvalues and we prove a localization result about the associated eigenfunctions. We also estimate, for each chosen aperture, the number of eigenvalues below the threshold of the essential spectrum.

La théorie des matrices s'est développée rapidement au cours des dernières décennies en raison de son large éventail d'applications et de ses nombreux liens avec différents domaines des mathématiques, de l'économie, de l'apprentissage automatique et du traitement du signal. Cette thèse concerne trois axes principaux liés à deux objets d'étude fondamentaux de la théorie des matrices et apparaissant naturellement dans de nombreuses applications, à savoir les matrices semi-définies positives et les matrices doublement stochastiques.Un concept qui découle naturellement du domaine de l'apprentissage automatique et qui est lié à la propriété semi-définie positive est celui des matrices de similarité. En fait, les matrices de similarité qui sont semi-définies positives revêtent une importance particulière en raison de leur capacité à définir des distances métriques. Cette thèse explorera la propriété semi-définie positive pour une liste de matrices de similarité trouvées dans la littérature. De plus, nous présentons de nouveaux résultats concernant les propriétés définie positive et semi-définie trois-positive de certains matrices de similarité. Une discussion détaillée des nombreuses applications de tous ces propriétés dans divers domaines est également établie.D'autre part, un problème récent de l'analyse matricielle implique l'étude des racines des matrices stochastiques, ce qui s'avère important dans les modèles de chaîne de Markov en finance. Nous étendons l'analyse de ce problème aux matrices doublement stochastiques semi-définies positives. Nous montrons d'abord certaines propriétés géométriques de l'ensemble de toutes les matrices semi-définies positives doublement stochastiques d'ordre n ayant la p-ième racine doublement stochastique pour un entier donné p . En utilisant la théorie des M-matrices et le problème inverse des valeurs propres des matrices symétriques doublement stochastiques (SDIEP), nous présentons également quelques méthodes pour trouver des classes de matrices semi-définies positives doublement stochastiques ayant des p-ièmes racines doublement stochastiques pour tout entier p.Dans le contexte du SDIEP, qui est le problème de caractériser ces listes de nombres réels qui puissent constituer le spectre d’une matrice symétrique doublement stochastique, nous présentons quelques nouveaux résultats le long de cette ligne. En particulier, nous proposons d’utiliser une méthode récursive de construction de matrices doublement stochastiques afin d'obtenir de nouvelles conditions suffisantes indépendantes pour SDIEP. Enfin, nous concentrons notre attention sur les spectres normalisés de Suleimanova, qui constituent un cas particulier des spectres introduits par Suleimanova. En particulier, nous prouvons que de tels spectres ne sont pas toujours réalisables et nous construisons trois familles de conditions suffisantes qui affinent les conditions suffisantes précédemment connues pour SDIEP dans le cas particulier des spectres normalisés de Suleimanova. / Matrix theory has shown its importance by its wide range of applications in different fields such as statistics, machine learning, economics and signal processing. This thesis concerns three main axis related to two fundamental objects of study in matrix theory and that arise naturally in many applications, that are positive semi-definite matrices and doubly stochastic matrices.One concept which stems naturally from machine learning area and is related to the positive semi-definite property, is the one of similarity matrices. In fact, similarity matrices that are positive semi-definite are of particular importance because of their ability to define metric distances. This thesis will explore the latter desirable structure for a list of similarity matrices found in the literature. Moreover, we present new results concerning the strictly positive definite and the three positive semi-definite properties of particular similarity matrices. A detailed discussion of the many applications of all these properties in various fields is also established.On the other hand, an interesting research field in matrix analysis involves the study of roots of stochastic matrices which is important in Markov chain models in finance and healthcare. We extend the analysis of this problem to positive semi-definite doubly stochastic matrices.Our contributions include some geometrical properties of the set of all positive semi-definite doubly stochastic matrices of order n with nonnegative pth roots for a given integer p. We also present methods for finding classes of positive semi-definite doubly stochastic matrices that have doubly stochastic pth roots for all p, by making use of the theory of M-Matrices and the symmetric doubly stochastic inverse eigenvalue problem (SDIEP), which is also of independent interest.In the context of the SDIEP, which is the problem of characterising those lists of real numbers which are realisable as the spectrum of some symmetric doubly stochastic matrix, we present some new results along this line. In particular, we propose to use a recursive method on constructing doubly stochastic matrices from smaller size matrices with known spectra to obtain new independent sufficient conditions for SDIEP. Finally, we focus our attention on the realizability by a symmetric doubly stochastic matrix of normalised Suleimanova spectra which is a normalized variant of the spectra introduced by Suleimanova. In particular, we prove that such spectra is not always realizable for odd orders and we construct three families of sufficient conditions that make a refinement for previously known sufficient conditions for SDIEP in the particular case of normalized Suleimanova spectra.