
A complex matrix A is said to be nonnormal if it satisfies the equation AA^⊤ + A^⊤A = 0 . This paper gives the derivation of certain classes of nonnormal Toeplitz matrices.
Using the Stefan problem as an example, the validity of a modified variational formulation of the discontinuous Galerkin method in a three-dimensional domain with a moving boundary is investigated. The h-adaptivity strategy is applied to solve this class of problems, which leads to a large size of the discrete analog. The new idea in the finite element approximation is to apply the principle of decomposing the numerical solution into a continuous macroscale component and a discontinuous microscale component near the phase transition front, introducing a special operator linking the two scales into the variational formulation. This strategy addresses two main problems arising in the discretization of problems in domains with moving boundaries. The first problem is related to constructing a weak solution in the space of L2-class functions and preserving the global regularity property of the mathematical model at the discrete level. For this purpose, a locally conservative computational scheme of the discontinuous Galerkin method is used, which makes it possible to obtain physically relevant results in domains with a rapidly varying solution gradient. The second problem is the explosive growth in the number of degrees of freedom when using the discontinuous Galerkin computational scheme throughout the entire computational domain. Since the microscale discontinuous component of the solution is introduced locally, the discrete analog of the Stefan problem has a significantly smaller size. The article presents a numerical analysis of the spectral properties of discrete analogs, the projection operator linking the large-scale (continuous) and microscale (discontinuous) function spaces, as well as the results of numerical simulation of the ice melting process in a three-dimensional medium with a moving boundary.
A control problem for a quasilinear heat transfer equation with multiplicative and distributed controls is studied. Based on the analysis of an optimality system, the bang-bang property of distributed control is established and stability estimates of small optimal solutions with respect to small perturbations of the cost functional are derived.
The paper focuses on a problem of propagation of a superposition of two electromagnetic waves in a planar shielded waveguide filled with an anisotropic nonhomogeneous nonlinear medium. The physical problem for Maxwell’s equations is reduced to a nonlinear two-parameter eigenvalue problem. A numerical method is suggested for a fairly general formulation of the problem. For an important special case of the problem, an analytical approach is developed allowing one to prove the existence of solutions, including nonlinearizable ones. Numerical results are also presented.
We consider the potential theory for the operator of the linear part of a certain model nonlinear equation. This theory is substantially used in the second part of the study, where we investigate the qualitative properties of the solution to the Cauchy problem for the corresponding nonlinear equation.