In this thesis, we consider second-order fully nonlinear partial differential equations of elliptic type. Our aim is to develop computational methods using convergent difference schemes for stationary Hamilton-Jacobi equations with Dirichlet and Neumann type boundary conditions in arbitrary two-dimensional domains. First, we introduce the notion of viscosity solutions in both continuous and discontinuous frameworks. Next, we review Barles-Souganidis approach using monotone, consistent, and stable schemes. In particular, we show that these schemes converge locally uniformly to the unique viscosity solution of the first-order Hamilton-Jacobi equations under mild assumptions. To solve the scheme numerically, we use Euler map with some initial guess. This iterative method gives the viscosity solution as a limit. Moreover, we illustrate our numerical approach in several two-dimensional examples.

[pt] No desenvolvimento das técnicas de determinação do campo elétrico irradiado por refletores simétricos de grandes dimensões, vem sendo constante a preocupação dos pesquisadores com o aprimoramento de métodos numéricos que melhor e mais rapidamente possam gerar dados confiáveis da radiação destas antenas, tanto na região de campo-distante quanto na região de Fresnel. Dentre as técnicas de cálculo possíveis, a utilização das séries de Jacobi-Bessel destaca-se como uma ferramenta bastante útil para a determinação do diagrama de radiação paraxial (na região próxima ao eixo de simetria do refletor) nas regiões de Fresnel e de Campo-Distante. Esta técnica é certamente conveniente quando é necessário calcular o campo para um grande número de pontos de observação. O presente trabalho contém, além do desenvolvimento teórico detalhado do método em questão para a determinação do campo espalhado por parabolóides simétricos alimentados no foco, aplicações nas quais são comparados os resultados com os obtidos através da aplicação de métodos tradicionais tais como: Integração Direta de Correntes, G.T.D. e Óitca Geométrica. São apresentadas e analisadas algumas curvas representativas da convergência das séries em campo próximo (Fresnel) e distante e para diversas situações do sistema (dimensões da antena, freqüência e ponto de observação). É feita ainda uma análise crítica da validade das aproximações de Fresnel na integral de Radiação [1] e das aproximações F.S.A. (Frasnel Small Angle Approximations) introduzidas no desenvolvimento das séries de Jacobi-Bessel. Finalmente, um programa de computador em linguagem FORTRAN desenvolvido para o cálculo do espalhamento de parabolóides simétricos On-Axis usando as séries de Jacobi-Bessel, é apresentado em anexo. / [en] This thesis reports on the computation of radiation patterns of axially symmetric parabolic reflector antennas using the Jacobi-Bessel Series Method. A detailed presentation of the method for the computation in the near far field regions is followed by a discussion about the validity of the approximations involved. Examples of application are given and the results, when compared with P.O., G.O. and solutions show good agreement. The ability of the method to achieve convergence faster than the classical current integration method is stressed. A complete listing of a fully commented FORTRAN IV computer program, developed by the author, is also presented.

<div>We develop a growth option and asset pricing model that incorporates uncertain cash flow volatility by way of a bounded quadratic diffusion. Using different measures of risk uncertainty, we study the combined effects of risk and its associated uncertainty on project values, firm investment, and the resulting returns. Uncertain cash flow volatility is modeled by a Jacobi process, and our main interest is the effect of the max uncertainty arising from the diffusion term. For comparison, we also model the volatility by a CIR process. In regards to the Jacobi process, we consider upper and lower bounds on cash flow volatility as measures of uncertainty. For the max uncertainty and upper bound, we find that higher uncertainty leads to less investment, higher returns, and lower project values. In the case of the lower bound, we find that higher uncertainty leads to more investment, lower returns, and higher project values. Comparatively, using a CIR process in place of the Jacobi process yields differences in returns and growth option values, showing the importance of the diffusion term in the volatility process. Finally, we have reduced the computational complexity of the simulation. This allows the user to generate long time series and run cross sectional regressions with many firms.</div>

Cette thèse s’intéresse au traitement de la notion de vie au sein du système hégélien. Que ce soit comme vie logique, organique ou comme vie de l’esprit, cette notion représente un élément structurant dans le discours philosophique de Hegel. À un point tel que l’on peut, à partir de la systématicité dynamique de la vie et du vivant tel qu’ils sont thématisés chez Hegel, penser le caractère dynamique et vivant du système hégélien. Ce faisant, nous sommes en mesure de situer l’entreprise hégélienne dans le contexte de la philosophie postkantienne visant à un renouvellement de la pratique philosophique comme philosophie vivante capable d’articuler raison spéculative et vie pratique. Cela permet également de mieux situer l’entreprise hégélienne vis-à-vis des critiques tendant à faire de son système un système clos et ossifié. Partant du système hégélien tel qu’il se présente au cours de la période berlinoise au sein de l’Encyclopédie des sciences philosophiques, nous retraçons ainsi l’entreprise hégélienne dans sa volonté de penser à la fois la vie de la pensée et la rationalité du vivant.

Mémoire numérisé par la Direction des bibliothèques de l'Université de Montréal.

No description available.

I present two differential-algebraic (DA) methods for approximately solving the Hamilton- Jacobi (HJ) equation. I use the “automatic differentiation” property of DA to convert the nonlinear partial-differential HJ equation into a initial-value problem for a DA-valued first-order ordinary differential equation (ODE), the “HJ/DA equation”. The solution of either form of the HJ/DA equation is equivalent to a perturbative expansion of Hamilton’s principle function about some reference trajectory (RT) through the system. The HJ/DA method also extracts the equations of motion for the RT itself. Hamilton’s principle function generates the canonical transformation, or mapping, between the initial and final state of every trajectory through the system. Since the map is represented by a generating function, it must automatically be symplectic, even in the presence of round-off error. The DA-valued ODE produced by either form of HJ/DA is equivalent tc a hierarchically-ordered system of real-valued ODEs without “feedback” terms; therefore the hierarchy may be truncated at any (arbitrarily high) order without loss of self consistency. The HJ/DA equation may be numerically integrated using standard algorithms, if all mathematical operations are done in DA. I show that the norm of the DA-valued part of the solution is bounded by linear growth. The generating function may be used to track either particles or the moments of a particle distribution through the system. In the first method, all information about the perturbative dynamics is contained in the DA-valued generating function. I numerically integrate the HJ/DA equation, with the identity as the initial generating function. A difficulty with this approach is that not all canonical transformations can be represented by the class of generating functions connected to the identity; one finds that with the required initial conditions, the generating function becomes singular near caustics or foci. One may continue integrating through a caustic by using a Legendre transformation to obtain a new (but equivalent) generating function which is singular near the identity, but nonsingular near the caustic. However the Legendre transformation is a numerically costly procedure, so one would not want to do this often. This approach is therefore not practical for systems producing periodic motions, because one must perform a Legendre transformation four times per period. The second method avoids the caustic problem by representing only the nonlinear part of the dynamics by a generating function. The linearized dynamics is treated separately via matrix techniques. Since the nonlinear part of the dynamics may always be represented by a near-identity transformation, no problem occurs when passing through caustics. I successfully verify the HJ/DA method by applying it to three problems which can be solved in closed form. Finally, I demonstrate the method’s utility by using it to optimize the length of a lithium lens for minimum beam divergence via the moment-tracking technique. / Ph. D.

Alors qu'initialement utilisés pour résoudre le Problème du Logarithme Discret (DLP) dans le groupe de points d'une courbe elliptique, les couplages sont très à la mode en cryptographie ces années car ils permettent de construire de nouveaux protocoles cryptographiques. Cependant, le calcul efficace du couplage dépend de l'arithmétique du modèle de courbe elliptique choisi et du corps sur lequel cette courbe est définie. Dans cette thèse, nous calculons le couplage sur deux modèles de Jacobi de courbes elliptiques puis nous introduisons et étudions l'arithmétique d'un nouveau modèle d'Ewards de courbe elliptique défini en toutes caractéristiques. Plus précisément, Nous utilisons l'interprétation géométrique de la loi de groupe sur l'intersection des quadriques de Jacobi pour obtenir pour la première fois dans la littérature, les formules explicites de la fonction de Miller pour le calcul du couplage de Tate sur cette courbe. Pour un calcul de couplage avec un degré de plongement pair, nous définissons la tordue quadratique pour obtenir des étapes de doublement et d'addition efficaces dans l'algorithme de Miller. Ensuite nous utilisons un isomorphisme entre la quartique spéciale de Jacobi Ed: Y²=dX⁴+Z⁴ et le modèle de Weierstrass pour obtenir la fonction de Miller nécessaire au calcul du couplage de Tate. Pour un degré de plongement divisible par 4, nous définissons la tordue d'ordre 4 de cette courbe pour obtenir un résultat meilleur du calcul du couplage de Tate par rapport aux courbes elliptiques sous forme de Weierstrass. Notre résultat améliore en même temps les derniers résultats obtenus sur cette courbe. Ce résultat est donc le meilleur connu à ce jour, à notre connaissance, pour le calcul du couplage de Tate sur les courbes possédant des tordues d'ordre 4. En 2006, Hess et al. introduisent le couplage Ate, qui est une version améliorée du couplage de Tate. Nous calculons ce couplage et ses variantes sur la même quartique. Nous y obtenons encore des résultats meilleurs. Notre troisième contribution est l'introduction d'un nouveau modèle d'Edwards de courbe elliptique d'équation 1+x²+y²+x²y²=Xxy. Ce modèle est ordinaire sur les corps de caractéristique 2 et nous montrons qu'il est birationnellement équivalent au modèle original d'Edwards x²+y²=c²(1+x²y²) en caractéristique différente de 2. Pour ce faire, nous utilisons la théorie des fonctions thêta et un modèle intermédiaire que nous appelons modèle thêta de niveau 4. Nous utilisons les relations de Riemann des fonctions thêta pour étudier l'arithmétique de ces deux courbes. Nous obtenons d'une part une loi de groupe complète, unifiée et en particulier compétitive en caractéristique 2 et d'autre part nous présentons les meilleures formules d'addition différentielle sur le modèle thêta de niveau 4. / While first used to solve the Discrete Logarithm Problem (DLP) in the group of points of elliptic curves, bilinear pairings are now useful to construct many public key protocols. The efficiency of pairings computation depends on the arithmetic of the model chosen for the elliptic curve and of the base field where the curve is defined. In this thesis, we compute and implement pairings on elliptic curves of Jacobi forms and we study the arithmetic of a new Edwards model for elliptic curves defined over any finite field. More precisely, We use the geometric interpretation of the group law of Jacobi intersection curves to obtain the first explicit formulas for the Miller function in Tate pairing computation in this case. For pairing computation with even embedding degree, we define and use the quadratic twist of this curve to obtain efficient formulas in the doubling and addition stages in Miller's algorithm. Moreover, for pairing computation with embedding degree divisible by 4 on the special Jacobi quartic elliptic curve Ed :Y²=dX⁴+Z⁴, we define and use its quartic twist to obtain a best result with respect to Weierstrass curves. Our result is at the same time an improvement of a result recently obtained on this curve, and is therefore, to our knowledge, the best result to date on Tate pairing computation among all curves with quartic twists. In 2006, Hess et al. introduced the concept of Ate pairing which is an improving version of the Tate pairing. We extend the computation of this pairing and its variations to the curve E_d. Again our theoretical results show that this curve offers the best performances comparatively to other curves with quartic twists, especially Weiertrass curves. As a third contribution, we introduce a new Edwards model for elliptic curves with equation 1+x²+y²+x²y²=\lambda xy. This model is ordinary over binary fields and we show that it is birationally equivalent to the well known Edwards model x²+y²=c²(1+x²y²) over non-binary fields. For this, we use the theory of theta functions to obtain an intermediate model that we call the level 4 theta model. We study the arithmetic of these curves, using Riemann relations of theta functions. The group laws are complete, unified, efficient and are particularly competitive in characteristic 2. Our formulas for differential addition on the level four theta model over binary fields are the best to date among well known models of elliptic curves.

Ce travail porte sur la modélisation et la simulation du trafic routier sur un réseau. Modéliser le trafic sur une section homogène (c'est-à-dire sans entrée, ni sortie) trouve ses racines au milieu du XXème siècle et a généré une importante littérature depuis. Cependant, la prise en compte des discontinuités des réseaux comme les jonctions, n'a attiré l'attention du cercle scientifique que bien plus récemment. Pourtant, ces discontinuités sont les sources majeures des congestions, récurrentes ou non, qui dégradent la qualité de service des infrastructures. Ce travail se propose donc d'apporter un éclairage particulier sur cette question, tout en s'intéressant aux problèmes d'échelle et plus particulièrement au passage microscopique-macroscopique dans les modèles existants. La première partie de cette thèse est consacrée au lien existant entre les modèles de poursuite microscopiques et les modèles d'écoulement macroscopiques. Le passage asymptotique est assuré par une technique d'homogénéisation pour les équations d'Hamilton-Jacobi. Dans une deuxième partie, nous nous intéressons à la modélisation et à la simulation des flux de véhicules au travers d'une jonction. Le modèle macroscopique considéré est bâti autour des équations d'Hamilton-Jacobi. La troisième partie enfin, se concentre sur la recherche de solutions analytiques ou semi-analytiques, grâce à l'utilisation de formules de représentation permettant de résoudre les équations d'Hamilton-Jacobi sous de bonnes hypothèses. Nous nous intéressons également dans cette thèse, à la classe générique des modèles macroscopiques de trafic de second ordre, dits modèles GSOM / This work focuses on modeling and simulation of traffic flows on a network. Modeling road traffic on a homogeneous section takes its roots in the middle of XXth century and it has generated a substantial literature since then. However, taking into account discontinuities of the network such as junctions, has attracted the attention of the scientific circle more recently. However, these discontinuities are the major sources of traffic congestion, recurring or not, that basically degrades the level of service of road infrastructure. This work therefore aims to provide a unique perspective on this issue, while focusing on scale problems and more precisely on microscopic-macroscopic passage in existing models. The first part of this thesis is devoted to the relationship between microscopic car-following models and macroscopic continuous flow models. The asymptotic passage is based on a homogenization technique for Hamilton-Jacobi equations. In a second part, we focus on the modeling and simulation of vehicular traffic flow through a junction. The considered macroscopic model is built on Hamilton-Jacobi equations as well. Finally, the third part focuses on finding analytical or semi-analytical solutions, through representation formulas aiming to solve Hamilton-Jacobi equations under adequate assumptions. In this thesis, we are also interested in a generic class of second order macroscopic traffic flow models, the so-called GSOM models

Neste trabalho buscamos caracterizar o espectro de uma classe de operadores bloco--Jacobi limitados definidos em $l^2(\\Lambda,\\mathbb{C}^L)$ ($\\Lambda: \\mathbb{Z}_+\\times\\{0,1,\\ldots,L-1\\}$ representa uma faixa de largura $L\\ge 2$ no semi--plano $\\mathbb{Z}_+^2$) e sujeitos a perturbações esparsas (no sentido que as distâncias entre as ``barreiras\'\' crescem geometricamente à medida que estas se afastam da origem) distribuídas aleatoriamente. Tais operadores são construídos a partir da soma de Kronecker de matrizes de Jacobi $J$, cada qual atuando em uma direção do espaço. Demonstramos, por meio da bloco--diagonalização do operador, que %o estudo de suas principais propriedades espectrais dependem da %se limita à caracterização da ``medida de mistura\'\' $\\frac{1}{L}\\sum_{j=0}^{L-1}\\mu_j$, $\\mu_j$ a medida espectral associada à matriz de Jacobi $J^j=J+2\\cos(2\\pi j/L)I $. Para tanto, buscamos primeiramente caracterizar cada uma das medidas $\\mu_j$, explorando e aperfeiçoando algumas técnicas bastante conhecidas no estudo de operadores esparsos unidimensionais. Demonstramos, por exemplo, que a seqüência de ângulos de Prüfer (variáveis que, juntamente com os raios de Prüfer, parametrizam as soluções da equação de autovalores) é uniformemente distribuída no intervalo $[0,\\pi)$, o %que %resultado que nos permite determinar o comportamento assintótico médio das soluções da equação de autovalores. Tal resultado, aliado às técnicas desenvolvidas por Marchetti \\textit{et. al.} em \\cite{MarWre} e a uma adaptação dos critérios de Last e Simon \\cite{LS} para operadores esparsos, nos permitem demonstrar a existência de uma transição aguda (pontual) entre os espectros singular--contínuo e puramente pontual. Empregamos em seguida os resultados de Jitomirskaya e Last presentes em \\cite{JitLast} e obtemos a dimensão Hausdorff exata associada à medida $\\mu_j$, dada por $\\alpha_j=1+\\frac{4(1-p^2)^2}{p^2(4- (\\lambda-2\\cos(2\\pi j/L))^2)}$ ($\\lambda\\in[-2,2]$), recuperando um resultado análogo obtido por Zlato\\v s em \\cite{Zla}. Por fim, adaptamos tais resultados à situação da medida de mistura associada à matriz bloco--Jacobi, obtendo $\\alpha=\\min_{j\\in\\mathcal{I}(\\lambda)}\\alpha_j$, $\\mathcal{I}(\\lambda):\\{m \\in\\{0,1,\\ldots,L-1\\}:\\lambda\\in[-2+2\\cos(2\\pi j/L),2+2\\cos(2\\pi j/L)]\\}$, como sua dimensão Hausdorff exata. Estudamos modelos idênticos com esparsidades sub e super-geométricas, obtendo na primeira situação um espectro puramente pontual (de dimensão Hausdorff nula) e na segunda um espectro puramente singular--contínuo (de dimensão Hausdorff 1). Finalmente, verificamos a existência de transição entre os espectros puramente pontual e singular--contínuo em um modelo com esparsidade super-geométrica cuja dimensão Hausdorff associada à medida espectral é nula. / In this work we attempt to caracterize the spectrum of a class of limited block--Jacobi operators defined in $l^2(\\Lambda,\\mathbb{C}^L)$ ($\\Lambda: \\mathbb{Z}_+\\times\\{0,1,\\ldots,L-1\\}$ represents a strip of width $L\\ge 2$ on the semi--plane $\\mathbb{Z}_+^2$) subject to a sparse perturbation (which means that the distance between the ``barries\'\' grow geometrically with their distance to the origin) randomly distributed. Such operators are defined as Kronecker sums of unidimensional Jacobi matrices $J$, each one acting in different directions of the space. We prove, by means of a block--diagonalization of the operator, that %the study of its most relevant spectral properties depend on %is related to the caracterization of the ``mixture measure\'\' $\\frac{1}{L}\\sum_{j=0}^{L-1}\\mu_j$, $\\mu_j$ the spectral measure of the Jacobi matrix $J^j=J+2\\cos(2\\pi j/L)I$. For this, we must characterize at first each one of the measures $\\mu_j$, exploiting and improving some well known techniques developed in the study of unidimensional sparse operators. We prove, for instance, that the sequence of Prüfer angles (variables which parametrize the solutions of the eigenvalue equation) are uniform distributed on the interval $[0,\\pi)$, a result which gives us condition to determine the average asymptotic behavior of the solutions of the eigenvalue equation. Such result, in association with the techniques developed by Marchetti \\textit{et. al.} in \\cite{MarWre} and with an adaptation of Last--Simon \\cite{LS} criteria for sparse operator, permit us to prove the existence of a sharp transition between singular continuous and pure point spectra. Following on, we use the results from Jitomirskaya--Last of \\cite{JitLast} and obtain the exact Hausdorff dimension of the measure $\\mu_j$, given by $\\alpha_j=1+\\frac{4(1-p^2)^2}{p^2(4-(\\lambda-2\\cos(2\\pi j/L))^2)}$ ($\\lambda\\in[- 2,2]$), recovering an analogous result due to Zlato\\v s in \\cite{Zla}. At last, we adapt these results to the mixture measure of the block--Jacobi matrix, obtaining $\\alpha=\\min_{j\\in\\mathcal{I}(\\lambda)}\\alpha_j$, $\\mathcal{I}(\\lambda):\\{m \\in\\{0,1,\\ldots,L-1\\}:\\lambda\\in[-2+2\\cos(2\\pi j/L),2+2\\cos(2\\pi j/L)]\\}$, as its exact Hausdorff dimension. We study as well identical models with sub and super geometric sparsities conditions, obtaining a pure point spectrum (with null Hausdorff dimension) in the first case, and a purely singular continuous spectrum (such that its Hausdorff dimension is 1) in the second. Finally, we prove the existence of a transition between pure point and singular continuous spectra in a model with sub--geometric sparsity whose Hausdorff dimension related to the spectral measure is null.