Made available in DSpace on 2014-10-09T12:31:37Z (GMT). No. of bitstreams: 0 / Made available in DSpace on 2014-10-09T14:00:43Z (GMT). No. of bitstreams: 1 12885.pdf: 941021 bytes, checksum: 89af8f7af86b94c2e39f9dcd1cdcfa0b (MD5) / Dissertacao (Mestrado) / IEA/D / Escola Politecnica, Universidade de Sao Paulo - POLI/USP

Coordinated missions with platoons of autonomous agents are rapidly becoming popular because of technological advances in computing, networking, miniaturization and combination of electromechanical systems. These multi-agents networks coordinate their actions to perform challenging spatially-distributed tasks such as search, survey, exploration, and mapping. Environmental monitoring and locational optimization are among the main applications of the emerging technology of wireless sensor networks where the optimality refers to the assignment of sub-regions to each agent, in such a way that a suitable coverage metric is maximized. Usually the coverage metric encodes a distribution of risk defined on the area, and a measure of the performance of individual robots with respect to points inside the region of interest. The risk density can be used to quantify spatial distributions of risk in the domain. The solution of the optimal control problem in which the risk measure is not time varying is well known in the literature, with the optimal con figuration of the robots given by the centroids of the Voronoi regions forming a centroidal Voronoi tessellation of the area. In other words, when the set of mobile robots converge to the corresponding centroids of the Voronoi tessellation dictated by the coverage metric, the coverage itself is maximized. In this work, it is considered a time-varying risk density evolving according to a diffusion equation with varying boundary conditions that quantify a time-varying risk on the border of the workspace. Boundary conditions model a time varying flux of external threats coming into the area, averaged over the boundary length, so that rather than considering individual kinematics of incoming threats it is considered an averaged, distributed effect. This approach is similar to the one commonly adopted in continuum physics, in which kinematic descriptors are averaged over spatial domain and suitable continuum fields are introduced to describe their evolution. By adopting a first gradient constitutive relation between the flux and the density, a simple diffusion equation is obtained. Asymptotic convergence and optimality of the non-autonomous system are studied by means of Barbalat's lemma and connections with varying boundary conditions are established. Some criteria on time-varying boundary conditions and evolution are established to guarantee the stabilities of agents' trajectories. A set of numerical simulations illustrate theoretical results.

Spatially discrete Nagumo equations have widespread physical applications, including modeling electrical impulses traveling through a demyelinated axon, an environment typical in multiple scle- rosis. We construct steady-state, single front solutions by employing a piecewise linear reaction term. Using a combination of Jacobi-Operator theory and the Sherman-Morrison formula we de- rive exact solutions in the cases of homogeneous and inhomogeneous diffusion. Solutions exist only under certain conditions outlined in their construction. The range of parameter values that satisfy these conditions constitutes the interval of propagation failure, determining under what circumstances a front becomes pinned in the media. Our exact solutions represent a very specific solution to the spatially discrete Nagumo equation. For example, we only consider inhomogeneous media with one defect present. We created an original script in MATLAB which algorithmically solves more general cases of the equation, including the case for multiple defects. The algorithmic solutions are then compared to known exact solutions to determine their validity.

For more than 100 years, the Navier-Stokes equations and various linearizations have been used as a model to study fluid dynamics. Recently, attention has been directed toward studying the nonnormality of linearized problems and developing convergent numerical schemes for simulation of these sytems. Numerical schemes for optimal control problems often require additional properties that may not be necessary for simulation; these properties can be critical when studying nonnormal problems. This research is concerned with approximating infinite dimensional optimal control problems with nonnormal system operators. We examine three different finite element methods for a specific convection-diffusion equation and prove convergence of the infinitesimal generators. Additionally, for two of these schemes, we prove convergence of the associated feedback gains. We apply these three schemes to control problems and compare the performance of all three methods. / Ph. D.

We develop a methodology for systematic inversion of quantitative stratigraphic models. Quantitative stratigraphic modeling predicts stratigraphy using numerical simulations of geologic processes. Stratigraphic inversion methodically searches the parameter space in order to detect models which best represent the observed stratigraphy. Model parameters include sea-level change, tectonic subsidence, sediment input rate, and transport coefficients. We successfully performed a fully automated process based stratigraphic inversion of a geologically complex synthetic model. Several one and two parameter inversions were used to investigate the coupling of process parameters. Source location and transport coefficient below base level indicated significant coupling, while the rest of the parameters showed only minimal coupling. The influence of different observable data on the inversion was also tested. The inversion results using misfit based on sparse, but time dependent sample points proved to be better than the misfit based on the final stratigraphy only, even when sampled densely. We tested several inversion schemes on the topography dataset obtained from the eXperimental EarthScape facility simulation. The clustering of model parameters in most of the inversion experiments showed the likelihood of obtaining a reasonable number of compatible models. We also observed the need for several different diffusion-coefficient parameterizations to emulate different erosional and depositional processes. The excellent result of the piecewise inversion, which used different parameterizations for different time intervals, demonstrate the need for development or incorporation of time-variant parameterizations of the diffusion coefficients. We also present new methods for applying boundary condition on simulation of diffusion processes using the finite-difference method. It is based on the straightforward idea that solutions at the boundaries are smooth. The new scheme achieves high accuracy when the initial conditions are non vanishing at the boundaries, a case which is poorly handled by previous methods. Along with the ease in implementation, the new method does not require any additional computation or memory. / Ph. D.

Le travail de cette thèse a porté sur l'étude mathématique et numérique de quelques modèles multi-échelles issus de la physique des matériaux. La première partie de ce travail est consacrée à l'homogénéisation mathématique d'un problème elliptique avec une petite échelle. Nous étudions le cas particulier d'un matériau présentant une structure périodique avec un défaut. En adaptant la théorie classique d'Avellaneda et Lin pour les milieux périodiques, on démontre qu'on peut approximer finement la solution d'un tel problème, notamment à l'échelle microscopique. Nous obtenons des taux de convergence dépendant de l'étalement du défaut. On démontre aussi quelques propriétés des fonctions de Green d'un problème elliptique périodique avec conditions de bord périodiques. Les dislocations sont des lignes de défaut de la matière responsables du phénomène de plasticité. Les deuxième et troisième parties de ce mémoire portent sur la simulation de dislocations, d'abord en régime stationnaire puis en régime dynamique. Nous utilisons le modèle de Peierls, qui couple échelle atomique et échelle mésoscopique. Dans le cadre stationnaire, on obtient une équation intégrodifférentielle non-linéaire avec un laplacien fractionnaire: l'équation de Weertman. Nous en étudions les propriétés mathématiques et proposons un schéma numérique pour en approximer la solution. Dans le cadre dynamique, on obtient une équation intégrodifférentielle à la fois en temps et en espace. Nous en faisons une brève étude mathématique, et comparons différents algorithmes pour la simuler. Enfin, dans la quatrième partie, nous étudions la limite macroscopique d'une chaîne d'atomes soumis à la loi de Newton. Des arguments formels suggèrent que celle-ci devrait être décrite par une équation des ondes non-linéaires. Or, nous démontrons --sous certaines hypothèses-- qu'il n'en est rien lorsque des chocs apparaissent / In this thesis we study mathematically and numerically some multi-scale models from materials science. First, we investigate an homogenization problem for an oscillating elliptic equation. The material under consideration is described by a periodic structure with a defect at the microscopic scale. By adapting Avellaneda and Lin's theory for periodic structures, we prove that the solution of the oscillating equation can be approximated at a fine scale. The rates of convergence depend upon the integrability of the defect. We also study some properties of the Green function of periodic materials with periodic boundary conditions. Dislocations are lines of defects inside materials, which induce plasticity. The second part and the third part of this manuscript are concerned with simulation of dislocations, first in the stationnary regime then in the dynamical regime. We use the Peierls model, which couples atomistic and mesoscopic scales and involves integrodifferential equations. In the stationary regime, dislocations are described by the so-called Weertman equation, which is nonlinear and involves a fractional Laplacian. We study some mathematical properties of this equation and propose a numerical scheme for approximating its solution. In the dynamical regime, dislocations are described by an equation which is integrodifferential in time and space. We compare some numerical methods for recovering its solution. In the last chapter, we investigate the macroscopic limit of a simple chain of atoms governed by the Newton equation. Surprisingly enough, under technical assumptions, we show that it is not described by a nonlinear wave equation when shocks occur

The boundary element method (BEM) has become a powerful method for the numerical solution of boundary-value problems (BVPs), due to its ability (at least for problems with constant coefficients) of reducing a BVP for a linear partial differential equation (PDE) defined in a domain to an integral equation defined on the boundary, leading to a simplified discretisation process with boundary elements only. On the other hand, the coefficients in the mathematical model of a physical problem typically correspond to the material parameters of the problem. In many physical problems, the governing equation is likely to involve variable coefficients. The application of the BEM to these equations is hampered by the difficulty of finding a fundamental solution. The first part of this thesis will focus on the derivation of the boundary integral equation (BIE) for the Laplace equation, and numerical results are presented for some examples using constant elements. Then, the formulations of the boundary-domain integral or integro-differential equation (BDIE or BDIDE) for heat conduction problems with variable coefficients are presented using a parametrix (Levi function), which is usually available. The second part of this thesis deals with the extension of the BDIE and BDIDE formulations to the treatment of the two-dimensional Helmholtz equation with variable coefficients. Four possible cases are investigated, first of all when both material parameters and wave number are constant, in which case the zero-order Bessel function of the second kind is used as fundamental solution. Moreover, when the material parameters are variable (with constant or variable wave number), a parametrix is adopted to reduce the Helmholtz equation to a BDIE or a BDIDE. Finally, when material parameters are constant (with variable wave number), the standard fundamental solution for the Laplace equation is used in the formulation. In the third part, the radial integration method (RIM) is introduced and discussed in detail. Modifications are introduced to the RIM, particularly the fact that the radial integral is calculated by using a pure boundary-only integral which relaxes the “star-shaped” requirement of the RIM. Then, the RIM is used to convert the domain integrals appearing in both BDIE and BDIDE for heat conduction and Helmholtz equations to equivalent boundary integrals. For domain integrals consisting of known functions the transformation is straightforward, while for domain integrals that include unknown variables the transformation is accomplished with the use of augmented radial basis functions (RBFs). The most attractive feature of the method is that the transformations are very simple and have similar forms for both 2D and 3D problems. Finally, the application of the RIM is discussed for the diffusion equation, in which the parabolic PDE is initially reformulated as a BDIE or a BDIDE and the RIM is used to convert the resulting domain integrals to equivalent boundary integrals. Three cases have been investigated, for homogenous, non-homogeneous and variable coefficient diffusion problems.

This thesis focuses on a certain type of nonlinear parabolic partial differential equations, i.e. PME and FDE. Chapter 1 consists of a survey on results related to PME and FDE, and a short review on some works about deriving gradient estimates in probabilistic ways. In Chapter 2 we estimate gradient on space variables of solutions to the heat equation on Euclidean space. The main idea is to construct two semimartingales by letting the solution and its gradient running backward on the path space of a diffusion process. Estimates derived from decompositions of those two semimartingales are then combined to give rise to an upper bound on gradient that only involves the maximum of the initial data and time variable. In particular, it is independent of the dimension. In Chapter 3 we carry the idea in Chapter 2 onto the study of positive solutions to PME or FDE, and obtained a similar type of bound on |∇u| for local solutions to PME or FDE on Euclidean space. In existing literature there have always been constraints on m. By considering a more general form of transformation on u and introducing a family of equivalent measures on path space, we add more flexibility to our method. Thus our result is valid for a larger range of m. For global solutions, when m violates our constraint, we need two-sided bound on u to control |∇u|. In Chapter 4 we utilize maximum principle to derive Li-Yau type gradient estimate for PME on a compact Riemannian manifold with Ricci curvature bounded from below. Our result is able to yield a Harnack inequality possessing the right order in time variable when the lower bound of Ricci curvature is negative.

Neste trabalho é desenvolvido um método, que utiliza funções de Green, para a solução analítica da equação de difusão de nêutrons em multigrupo, para um reator refletido, cujo fluxo tem dependência apenas radial e com distribuição de combustível não uniforme no cerne. As propriedades de moderação, difusão e absorção são consideradas diferentes no cerne e refletor. Uma distribuição de densidade de potência, que estabelece a condição de criticalidade do reator, é assumida a priori e determina a distribuição de combustível no cerne. Com auxílio das funções de Green e das condições de continuidade do fluxo e da densidade de corrente de nêutrons na interface cerne-refletor, a equação de difusão em multigrupo é transformada em um sistema de equações lineares, contendo como incógnitas os valores dos fluxos na interface entre as regiões. Resolvido esse sistema, obtém-se os valores dos fluxos na interface e, com eles, a distribuição de fluxo em cada região e para cada grupo. Como verificação do método proposto, é feita uma aplicação numérica, utilizando dois grupos de energia, para um reator TRIGA de 1MW. Nessa aplicação são calculadas, além das distribuições de fluxos para os dois grupos de energia, a distribuição de combustível no cerne, a massa crítica e a potência específica linear, para diferentes distribuições de densidade de potência. / In the present work a method is developed for applying Green\'s functions to obtain an analytical solution o£ the neutron diffusion equation to the case o£ a reflected reactor. The problem of a non-uniform fuel distribution in the core is treated. Multigroup theory is used and the neutron flux is assumed to have only radial dependence. Different values are employed to characterize the moderation, diffusion and absorption properties o£ the core and the reflector. A power density distribution which establishes the reactor critica1 condition \"a priori\" is assumed and is then used to calculate the fuel distribution. By using the Green\'s functions and the continuity relations (for neutron fluxes and neutron current densities) at the core-reflector interface, the multigroup diffusion equation is transformed into a system of linear equations. In this system o£ equations the unknowns are the neutron fluxes at the core- reflector interface. Once this system is solved and the interface fluxes are determined, it follows immediately that the neutron flux distribution in the core and in the reflector is determined. The method employed and proposed in the present study has been applied to the problem of calculating the neutron distribution in a 1MW TRIGA reactor, using two energy group. This numerical application, in addition to calculating the two-group flux distribution, the fuel distribution in the core, the critical mass and the linear specific power for different assumed power density distribution have been evaluated.

Questões ambientais estão no centro das discussões nas últimas décadas. A poluição atmosférica, causada pela expansão pós-revolução industrial fez surgir a necessidade de aprender a descrever, usando modelos matemáticos, esse fenômeno. Com esse conhecimento pode-se propor soluções que mitiguem a poluição e os danos colaterais causados ao ambiente. A dispersão de poluentes modelada por soluções analíticas, a partir das equações de advecção-difusão oferecem um conhecimento sobre cada componente que constrói a equação, característica inexistente em outras abordagens, como a numérica. Entretanto ela era incapaz de descrever propriedades que se referem à turbulência, as estruturas coerentes, causadas por componentes não-lineares suprimidas por construção das equações governantes do modelo. Este trabalho estudou uma forma de recuperar características associadas à turbulência através de uma componente fundamental em estruturas coerentes, a fase. Essa é incluída no modelo que passa a descrever manifestações da turbulência em processos de dispersão através de flutuações de pequena escala na concentração da solução do modelo sesquilinear, que é determinístico-estocástico. No decorrer do trabalho há um estudo através de variações de parâmetros para compreender os efeitos da fase no modelo. Ele também foi aplicado ao experimento de Copenhagen e a dois cenários reais com a intenção de compreender o modelo frente à variáveis micrometeorológicas assim como aprimorá-lo para simular a dispersão de poluentes oriundos de fontes de forma realística. / Environmental issues have been at the center of discussions in the last few decades. Atmospheric pollution, caused by post-industrial revolution, has increased the necessity to describe, using mathematical models, this phenomenon. With this knowledge is possible to propose solutions mitigating the pollution and collateral damages caused in the environment. The pollutant dispersion modeled by analytical solutions, from advection-diffusion equations, offers a knowledge about each component that constructs the equation, a characteristic that does not exist in other approaches, such as numerical. However it was unable to describe properties that refer to turbulence, coherent structures, caused by nonlinear components suppressed by constructing the model governing equations. This work studied a way to recover characteristics associated with turbulence through a fundamental component in coherent structures, the phase. This is included in the model which describes manifestations of turbulence in the dispersion process through the presence of small-scale concentration fluctuations in the sesquilinear model, which is deterministicstochastic. In the course of this work there is a study through variations of parameters to understand the phase effects in the model. It was also applied to Copenhagen experiment and to two real scenarios with the intention of understanding the model regarding micrometeorological variables as well as improving it to simulate the pollutant dispersion from sources in a realistic way.