MSc (Mathematics) / Department of Mathematics

In the paper we construct exact three-point discretizations of linear and nonlinear two-point boundary value problems with boundary conditions of the first kind. The finite element approach uses basis functions defined by the coefficients of the differential equations. All the discretized boundary value problems are of inverse isotone type and so are its exact discretizations which involve tridiagonal M-matrices in the linear case and M-functions in the nonlinear case.

No description available.

No description available.

Many boundary value problems of mathematical physics are modelled by elliptic differential operators L in a given domain Ω . An effective method for treating such problems is the method of layer potentials, whose essence resides in reducing matters to solving a boundary integral equation. This, in turn, requires inverting a singular integral operator, naturally associated with L and Ω, on appropriate function spaces on ƌΩ. When the operator L is of second order and the domain Ω is Lipschitz (i.e., Ω is locally the upper-graph of a Lipschitz function) the fundamental work of B. Dahlberg, C. Kenig, D. Jerison, E. Fabes, N. Rivière, G. Verchota, R. Brown, and many others, has opened the door for the development of a far-reaching theory in this setting, even though several very difficult questions still remain unanswered. In this dissertation, the goal is to solve a number of open questions regarding spectral properties of singular integral operators associated with second and higher-order elliptic boundary value problems in non-smooth domains. Among other spectral results, we establish symmetry properties of harmonic classical double layer potentials associated with the Laplacian in the class of Lipschitz domains in R2. An array of useful tools and techniques from Harmonic Analysis, Partial Differential Equations play a key role in our approach, and these are discussed as preliminary material in the thesis: --Mellin Transforms and Fourier Analysis; --Calderón-Zygmund Theory in Uniformly Rectifiable Domains; -- Boundary Integral Methods. Chapter four deals with proving invertibility properties of singular integral operators naturally associated with the mixed (Zaremba) problem for the Laplacian and the Lamé system in infinite sectors in two dimensions, when considering their action on the Lebesgue scale of p integrable functions, for 1 < p < ∞. Concretely, we consider the case in which a Dirichlet boundary condition is imposed on one ray of the sector, and a Neumann boundary condition is imposed on the other ray. In this geometric context, using Mellin transform techniques, we identify the set of critical integrability indexes p for which the invertibility of these operators fails. Furthermore, for the case of the Laplacian we establish an explicit characterization of the Lp spectrum of these operators for each p є (1,∞), as well as well-posedness results for the mixed problem. In chapter five, we study spectral properties of layer potentials associated with the biharmonic equation in infinite quadrants in two dimensions. A number of difficulties have to be dealt with, the most significant being the more complex nature of the singular integrals arising in this 4-th order setting (manifesting itself on the Mellin side by integral kernels exhibiting Mellin symbols involving hyper-geometric functions). Finally, chapter six, deals with spectral issues in Lipschitz domains in two dimensions. Here we are able to prove the symmetry of the spectra of the double layer potentials associated with the Laplacian. This is in essence a two-dimensional phenomenon, as known examples show the failure of symmetry in higher dimensions. / Mathematics

Yes / A two-dimensional steady potential flow theory is applied to calculate the flow in an open channel with complex solid boundaries. The boundary integral equations for the problem under investigation are first derived in an auxiliary plane by taking the Cauchy integral principal values. To overcome the difficulties of a nonlinear curvilinear solid boundary character and free water surface not being known a priori, the boundary integral equations are transformed to the physical plane by substituting the integral variables. As such, the proposed approach has the following advantages: (1) the angle of the curvilinear solid boundary as well as the location of free water surface (initially assumed) is a known function of coordinates in physical plane; and (2) the meshes can be flexibly assigned on the solid and free water surface boundaries along which the integration is performed. This avoids the difficulty of the traditional potential flow theory, which seeks a function to conformally map the geometry in physical plane onto an auxiliary plane. Furthermore, rough bed friction-induced energy loss is estimated using the Darcy-Weisbach equation and is solved together with the boundary integral equations using the proposed iterative method. The method has no stringent requirement for initial free-water surface position, while traditional potential flow methods usually have strict requirement for the initial free-surface profiles to ensure that the numerical computation is stable and convergent. Several typical open-channel flows have been calculated with high accuracy and limited computational time, indicating that the proposed method has general suitability for open-channel flows with complex geometry.

Magnetic Resonance Electrical Impedance Mammography (MREIM) is a new imaging technique under development by Wollin Ventures, Inc. in conjunction with the H. Lee Moffitt Cancer Center & Research Institute. MREIM addresses the problem of low specificity of magnetic resonance mammography and high false-positive rates, which lead to unnecessary biopsies. Because cancerous tissue has a higher electrical conductivity than benign tissue, it may serve as a biomarker for differentiation between malignant and benign lesions. The MREIM principle is based on measuring both magnetic resonance and electric properties of the breast by adding a quasi-steady-state electric field to the standard magnetic resonance breast image acquisition. This applied electric field produces a current density that creates an additional magnetic field that in turn alters the native magnetic resonance signal in areas of higher electrical conductivity, corresponding to cancerous tissue. This work comprises MREIM theory, computer simulations, and experimental developments. First, a general overview and background review of tissue modeling and electrical-impedance imaging techniques are presented. The experimental part of this work provides a description of the MREIM apparatus and the imaging results of a custom-made breast phantom. This phantom was designed and developed to mimic the magnetic resonance and electrical properties of the breast. The theoretical part of this work provides an extension to the initial MREIM theoretical developments to further understand the MREIM effects. MREIM computer simulations were developed for both idealized and realistic tumor models. A method of numerical calculation of electric potential and induced magnetic field distribution in objects with irregular boundaries and anisotropic conductivity was developed based on the Finite Difference Method. Experimental findings were replicated with simulations. MREIM effects were analyzed with contrast diagrams to show the theoretical perceptibility as a function of the acquisition parameters. An important goal was to reduce the applied current. A new protocol for an MREIM sequence is suggested. This protocol defines parameters for the applied current synchronized to a specific magnetic resonance imaging sequence. A simulation utilizing this protocol showed that the MREIM effect is detectable for a 3-mm-diameter tumor with a current density of 0.5 A/m², which is within acceptable limits.

Numerical methods to solve initial value problems of differential equations progressed quite a bit in the last century. We give a brief summary of how useful numerical methods are for ordinary differential equations of first and higher order. In this thesis both computational and theoretical discussion of the application of numerical methods on differential equations takes place. The thesis consists of an investigation of various categories of numerical methods for the solution of ordinary differential equations including the numerical solution of ordinary differential equations from a number of practical fields such as equations arising in population dynamics and astrophysics. It includes discussion what are the advantages and disadvantages of implicit methods over explicit methods, the accuracy and stability of methods and how the order of various methods can be approximated numerically. Also, semidiscretization of some partial differential equations and stiff systems which may arise from these semidiscretizations are examined.

The paper deals with elliptic theory on manifolds with boundary represented as a covering space. We compute the index for a class of nonlocal boundary value problems. For a nontrivial covering, the index defect of the Atiyah-Patodi-Singer boundary value problem is computed. We obtain the Poincaré duality in the K-theory of the corresponding manifolds with singularities.

The aim of this paper is to explain the notion of subspace defined by means of pseudodifferential projection and give its applications in elliptic theory. Such subspaces are indispensable in the theory of well-posed boundary value problems for an arbitrary elliptic operator, including the Dirac operator, which has no classical boundary value problems. Pseudodifferential subspaces can be used to compute the fractional part of the spectral Atiyah–Patodi–Singer eta invariant, when it defines a homotopy invariant (Gilkey’s problem). Finally, we explain how pseudodifferential subspaces can be used to give an analytic realization of the topological K-group with finite coefficients in terms of elliptic operators. It turns out that all three applications are based on a theory of elliptic operators on closed manifolds acting in subspaces.