Refine
Year of publication
Language
- English (182)
Keywords
Project
Application Area
- D (182) (remove)
To approximate convolutions which occur in evolution equations with memory terms, a variable-stepsize algorithm is presented for which advancing $N$ steps requires only $O(N\log N)$ operations and $O(\log N)$ active memory, in place of $O(N^2)$ operations and $O(N)$ memory for a direct implementation. A basic feature of the fast algorithm is the reduction, via contour integral representations, to differential equations which are solved numerically with adaptive step sizes. Rather than the kernel itself, its Laplace transform is used in the algorithm. The algorithm is illustrated on three examples: a blow-up example originating from a Schrödinger equation with concentrated nonlinearity, chemical reactions with inhibited diffusion, and viscoelasticity with a fractional order constitutive law.
A new approach to derive transparent boundary conditions (TBCs) for wave, Schrödinger, heat and drift-diffusion equations is presented. It relies on the pole condition and distinguishes between physical reasonable and unreasonable solutions by the location of the singularities of the spatial Laplace transform of the exterior solution. To obtain a numerical algorithm, a Möbius transform is applied to map the Laplace transform onto the unit disc. In the transformed coordinate the solution is expanded into a power series. Finally, equations for the coefficients of the power series are derived. These are coupled to the equation in the interior, and yield transparent boundary conditions.
Numerical results are presented in the last section, showing that the error introduced by the new approximate TBCs decays exponentially in the number of coefficients.
In this review article we discuss different techniques to solve numerically the
time-dependent Schrödinger equation on unbounded domains.
We present in detail the most recent approaches and describe briefly alternative ideas pointing out the relations between these works.
We conclude with several numerical examples from
different application areas to compare the presented techniques. We mainly focus on the one-dimensional problem but also touch upon the situation in two space dimensions and the cubic nonlinear case.
We consider hybrid systems of differential-algebraic equations and present
a general framework for general nonlinear over- and underdetermined hybrid
systems that allows the
analysis of existence and uniqueness and the application of index reduction
methods for hybrid differential-algebraic systems.
A particular difficulty in the numerical simulation of hybrid systems is
(numerical) chattering, i.e., fast oscillations between modes of operations.
A regularization technique using sliding modes allows to regularize the
system behavior in the case of chattering.
Further, we show how chattering behavior during the numerical solution can
be prevented using sliding mode simulation. The advantage of the sliding mode
simulation is illustrated by numerical examples.
The nonlinear Schrödinger equation based on the Taylor approximation of the material dispersion can become invalid for ultrashort and few-cycle optical pulses. Instead, we use a rational fit to the dispersion function such that the resonances are naturally accounted for. This approach allows us to derive a simple non-envelope model for short pulses propagating in one spatial dimension. This model is further investigated numerically and analytically.
A new concept is introduced for the adaptive finite element discretization of partial differential equations that have a sparsely
representable solution. Motivated by recent work on compressed sensing, a recursive mesh refinement procedure is presented that uses linear programming to find a good approximation to the sparse solution on a given refinement level. Then only those parts of the mesh are refined that belong to nonzero expansion coefficients. Error estimates for this procedure are refined and the behavior of the procedure is demonstrated via some simple elliptic model problems.
We consider linear discrete-time descriptor systems, i.e., systems of linear equations of the form $E_{k+1} x^{k+1} = A_k x^k + f^k$ for $k \in \IZ$, where all $E_k$ and $A_k$ are matrices, $f_k$ are vectors and $x_k$ are the vectors of the solution we are looking for. We study the existence and uniqueness of solutions. A strangeness index is defined for such systems. Compared to the continuous-time case, it turns out, that in the discrete-time case it makes a difference, if one has an initial condition and one wants a solution in the future or if one has an initial condition and one wants a solution into the past and future at the same time.
Studying high-dimensional Hamiltonian systems with microstructure, it is an important and challenging problem to identify reduced macroscopic models that describe some effective dynamics on large spatial and temporal scales. This paper concerns the question how reasonable macroscopic Lagrangian and Hamiltonian structures can by derived from the microscopic system. In the first part we develop a general approach to this problem by considering non-canonical Hamiltonian structures on the tangent bundle. This approach can be applied to all Hamiltonian lattices (or Hamiltonian PDEs) and involves three building blocks: (i) the embedding of the microscopic system, (ii) an invertible two-scale transformation that encodes the underlying scaling of space and time, (iii) an elementary model reduction that is based on a Principle of Consistent Expansions. In the second part we exemplify the reduction approach and derive various reduced PDE models for the atomic chain. The reduced equations are either related to long wave-length motion or describe the macroscopic modulation of an oscillatory microstructure.
We study both theoretically and experimentally typical operation
regimes of 40 GHz monolithic mode-locked lasers. The underlying Traveling Wave Equation model reveals quantitative agreement for characteristics of the fundamental mode-locking as pulse width and repetition frequency tuning, as well as qualitative agreement with the experiments for other dynamic regimes. Especially the appearance of stable harmonic mode-locking at 80 GHz
has been predicted theoretically and confirmed by measurements.
Furthermore, we derive and apply a simplified Delay-Differential-Equation model
which guides us to a qualitative analysis of bifurcations responsible for the appearance
and the breakup of different mode-locking regimes. Higher harmonics of mode-locking are predicted by this model as well.
We explore the concept of passive-feedback lasers for direct signal
modulation at 40 Gbit/s. Based on numerical simulation and bifurcation
analysis, we explain the main mechanisms in these devices which are
crucial for modulation at high speed. The predicted effects are
demonstrated experimentally by means of correspondingly designed
devices. In particular a significant improvement of the modulation
bandwidth at low injection currents can be demonstrated.
We investigate a semiconductor laser with delayed optical feedback due
to an external cavity formed by a regular mirror. We discuss
similarities and differences of the well-known Lang-Kobayashi delay
differential equation model and the traveling wave partial
differential equation model. For comparison we locate the continuous
wave states in both models and analyze their stability.
We use the traveling wave model for simulating and analyzing
nonlinear dynamics of complex semiconductor ring laser devices.
This modeling allows to consider temporal-spatial distributions
of the coun\-ter-pro\-pa\-ga\-ting slowly varying optical fields
and the carriers, what can be important when studying
non-homogeneous ring cavities, propagation of short pulses or fast switching.
By performing numerical integration of the model
equations we observe several dynamic regimes as well as transitions
between them. The computation of ring cavity modes explains some
peculiarities of these regimes.
We propose a model reduction method for positive systems that ensures the positivity of the reduced-order model. In the standard as well as in the descriptor case, for continuous-time and discrete-time systems, our approach is based on constructing diagonal solutions of Lyapunov inequalities. These are linear matrix inequalities (LMIs), which are shown to be feasible. Positivity and stability are preserved and an error bound in the $\mathcal{H}_\infty$-norm is provided.
Three properties of matrices: the spark, the mutual incoherence and the restricted isometry property have recently been introduced in the context of compressed sensing. We study these properties for matrices that are Kronecker products and show how these properties relate to those of the factors. For the mutual incoherence we also
discuss results for sums of Kronecker products.
We prove an optimal regularity result for elliptic operators $-\nabla \cdot \mu \nabla:W^1,q_0 \rightarrow W^-1,q$ for a $q>3$ in the case when the coefficient function $\mu$ has a jump across a $C^1$ interface and is continuous elsewhere. A counterexample shows that the $C^1$ condition cannot be relaxed in general. Finally, we draw some conclusions for corresponding parabolic operators.
Let $\Upsilon$ be a three-dimensional Lipschitz polyhedron, and assume that the matrix function $\mu$ is piecewise constant on a polyhedral partition of $\Upsilon$. Based on regularity results for solutions to two-dimensional anisotropic transmission problems near corner points we obtain conditions on $\mu$ and the intersection angles between interfaces and $\partial \Upsilon$ ensuring that the operator $-\nabla \cdot \mu \nabla$ maps the Sobolev space $W^1,q_0(\Upsilon)$ isomorphically onto $W^-1,q(\Upsilon)$ for some $q > 3$.
Using a classical theorem of Sobolevskii on equations of parabolic type in a Banach space and recently obtained results on elliptic operators with discontinuous coefficients including mixed boundary conditions we prove that quasilinear parabolic systems in diagonal form admit a local, classical solution in the space of p-integrable functions, for some p greater than 1, over a bounded two dimensional space domain. As applications we have in mind systems of reaction diffusion equations, e.g. van Roosbroeck's system. The treatment of such equations in a space of integrable functions enables us to define the normal component of the flow across any part of the Dirichlet boundary by Gauss' theorem.
The paper is devoted to Schroedinger operators on bounded intervals of the real axis with dissipative boundary conditions. In the framework of the Lax-Phillips scattering theory the asymptotic behaviour of the phase shift is investigated in detail and its relation to the spectral shift is discussed, in particular, trace formula and Birman-Krein formula are verified directly. The results are used for dissipative Schroedinger-Poisson systems.
Electronic structure and optoelectronic properties of strained InAsSb/GaSb multi quantum-wells
(2008)
A study of the optical properties of a set of InAsxSb1-x/Al0.15In0.85As0.77Sb0.23/GaSb multiple quantum-wells (for x between 0.82 and 0.92) with build-in strains in the -0.62% to +0.05%-range is presented. The energy of the lowest quantum-confined optical transition is calculated by kp perturbation theory and experimentally determined by absorption measurements. Stokes shift of photoluminescence, photocurrent and of the emission from light emitting devices against the absorption edge of the quantum-well are quantified. The impact of the decreasing carrier confinement in the InAsxSb1-x quantum well system with increasing mole fraction is analyzed theoretically, and experimentally demonstrated by photoluminescence measurement. Our results allow for the improvement of optoelectronic devices, in particular for tailoring emission spectra of light emitting diodes.
We describe an embedding of a quantum mechanically described structure into a macroscopic flow. The open quantum system is partly driven by an adjacent macroscopic flow acting on the boundary of the bounded spatial domain designated to quantum mechanics. This leads to an essentially non-selfadjoint Schroedinger-type operator, the spectral properties of which will be investigated.
We study in detail Schroedinger-type operators on a bounded interval of the real axis with dissipative boundary conditions. The characteristic function of such operators is computed, its minimal self-adjoint dilation is constructed and the generalized eigenfunction expansion for the dilation is developed. The problem is motivated by semiconductor physics.
Non-selfadjoint operators play an important role in the modeling of open quantum systems. We consider a one-dimensional Schroedinger-type operator with dissipative boundary conditions and dissipative delta potentials. An explicit description of the characteristic function, the minimal dilation and the generalized eigenfunctions of the dilation is given. The quantities of carrier and current densities are rigorously defined. Furthermore we will show that the current is not constant and that the variation of the current depend essentially on the chosen density matrix and imaginary parts of the delta potentials. This correspondence can be used to model a recombination-generation rate in the open quantum system.
The classical singular value decomposition for a matrix $A\in\Cmn$ is a
canonical form for $A$ that also displays the eigenvalues
of the Hermitian matrices $AA^\ast$ and $A^\ast A$. In this paper, we develop
a corresponding decomposition for $A$ that provides the Jordan canonical forms
for the complex symmetric matrices $AA^T$ and $A^TA$. More generally, we consider
the matrix triple $(A,G_1,G_2)$, where $G_1\in\CC{m}, G_2\in\CC{n}$
are invertible and either complex symmetric and complex skew-symmetric, and we
provide a canonical form under transformations of the form
$(A,G_1,G_2)\mapsto(X^T A Y, X^T G_1X, Y^T G_2Y)$, where $X,Y$ are nonsingular.
We derive formulas for the minimal positive solution of a
particular non-symmetric Riccati
equation arising in transport theory. The formulas are based
on the eigenvalues of an
associated matrix. We use the formulas to explore some new
properties of the minimal positive solution and to derive
fast and highly accurate numerical methods. Some numerical tests
demonstrate the properties of the new methods.
Lyapunov and exponential dichotomy spectral theory is extended
from ordinary differential equations (ODEs) to nonautonomous
differential-algebraic equations (DAEs). By using orthogonal
changes of variables, the original DAE system is transformed into
appropriate condensed forms, for which concepts such as Lyapunov
exponents, Bohl exponents, exponential dichotomy and spectral
intervals of various kinds can be analyzed via the resulting
underlying ODE. Some essential differences between the spectral
theory for ODEs and that for DAEs are pointed out. Numerical
methods for computing the spectral intervals associated with
Lyapunov and Sacker-Sell (exponential dichotomy) spectra are
derived by modifying and extending those methods proposed for ODEs. Perturbation theory and error analysis are discussed, as
well. Finally, some numerical examples are presented to illustrate
the theoretical results and the properties of the numerical
methods.
The perturbation and ADAE index of a degenerated hyperbolic system modelling a heat exchanger
(2007)
The heat exchanger in a heat pump can be modelled by the zero Mach-number limit of the Euler equations of compressible fluid flow. This system turns out to be a coupled hyperbolic/parabolic equation with coupled, time-dependent boundary conditions. Using the theory of abstract differential-algebraic equations it is shown that the frozen coefficient system has ADAE index 1. Moreover, the much stronger result is proven that the system has time-perturbation index one and space-perturbation index two even in the case of time-dependent boundary conditions. The results are stated in terms of the original physical variables. The estimates agree well with numerical experiments.
We propose a model reduction method for positive systems that ensures the positivity of the reduced model. For both, continuous-time and discrete-time systems, our approach is based on constructing diagonal solutions of Lyapunov inequalities. These are linear matrix inequalities (LMIs), which are shown to be feasible. Stability is preserved and an error bound in the $\mathcal{H}_\infty$-norm is provided.
The purpose of this paper is the analysis of dynamic iteration methods for
the numerical integration of coupled systems of ODEs and DAEs.
We will investigate convergence of these methods and put special emphasis
on the {\sc Jacobi}- and {\sc Gauss-Seidel} methods. Furthermore, the
fundamental difference in the convergence behaviour of coupled ODEs and DAEs
is pointed out. This difference is used to explain why certain relaxation methods
for coupled DAEs may fail. Finally, a remedy to this undesirable
effect is proposed that makes use of a so-called {\em preconditioned dynamic
iteration} strategy. This regularization also allows significant reduction of
dynamic iteration steps.
In this paper, we empirically investigate the NP-hard problem of finding sparsest solutions to linear equation systems, i.e., solutions with as few nonzeros as possible. This problem has received considerable interest in the sparse approximation and signal processing literature, recently. We use a branch-and-cut approach via the maximum feasible subsystem problem to compute optimal solutions for small instances and investigate the uniqueness of the optimal solutions. We furthermore discuss five (modifications of) heuristics for this problem that appear in different parts of the literature. For small instances, the exact optimal solutions allow us to evaluate the quality of the heuristics, while for larger instances we compare their relative performance. One outcome is that the so-called basis pursuit heuristic performs worse, compared to the other methods. Among the best heuristics are a method due to Mangasarian and a bilinear approach.
In this paper we consider structure-preserving model reduction of
second-order systems using a~ba\-lan\-ced truncation approach.
Several sets of singular values are introduced for such systems,
which lead to different concepts of balancing and different
second-order balanced truncation methods. We compare the
properties of these methods on numerical examples.
In this paper we consider the rational interpolation problem consisting in finding a rational matrix-valued function that
interpolates a given set of parameters. We briefly describe two different numerical methods for solving this problem. These are
the vector fitting and the frequency domain subspace identification method. Several numerical examples are given that compare
the properties of these methods. Furthermore, we discuss the computation of a (minimal) state space realization of a rational
function. Model order reduction methods such as modal approximation and balanced truncation are also presented. These
methods can be used to compute a reduced-order approximation of the realized dynamical system.
Passivation of LTI systems
(2007)
In this paper we consider a passivation procedure for linear time-invariant systems.
This procedure is based on the spectral properties of related Hamiltonian matrices.
We also present a structure-preserving algorithm for computing the imaginary
eigenvalues and the corresponding eigenvectors of Hamiltonian matrices.
Numerical examples are given.
Perturbation of Purely Imaginary Eigenvalues of Hamiltonian Matrices under Structured Perturbations
(2007)
We discuss the perturbation theory for purely imaginary eigenvalues of Hamiltonian matrices under Hamiltonian and non-Hamiltonian perturbations. We
show that there is a substantial difference in the behavior under these perturbations. We also discuss the perturbation of real eigenvalues of real
skew-Hamiltonian matrices under structured perturbations and use these results to analyze the properties of the URV method of computing the
eigenvalues of Hamiltonian matrices.
We review some known results for POD model reduction applied to ODEs. Then, these
results are generalized to several types of DAEs. We provide algorithms for the
model reduction and error bounds for the reduced order models. Some limits of
the approach are pointed out and alternative methods for reduced order subspace approximation
are presented. The POD approach is tested and evaluated for a medium sized DAE example
from multibody dynamics.
Classical stability properties of solutions
that are well-known for ordinary differential
equations (ODEs) are generalized to differential-algebraic equations (DAEs).
A new test equation is derived for the analysis of numerical methods applied
to DAEs with respect to the stability of the numerical approximations.
Morevover, a stabilization technique is developed to improve the stability of classical DAE integration methods. The stability regions for these stabilized discretization methods are determined and it is shown that they much better reproduce the stability properties known for the ODE case
than in the unstabilized form.
Movies that depict the stability regions for several methods are included for interactive use.
We study optimal control problems for general unstructured nonlinear differential-algebraic equations of arbitrary index.
In particular, we derive necessary conditions in the case of linear-quadratic control problems and extend them to the general nonlinear case.
We also present a Pontryagin maximum principle for general unstructured nonlinear DAEs in the case of restricted controls.
Moreover, we discuss the numerical solution of the resulting two-point boundary value problems and present a numerical example.
In this paper we introduce a new method for the computation of KKT matrices that arise from solving constrained, nonlinear optimization problems. This method requires updating of null-space factorizations after a low rank modification. The update procedure has the advantage that it is significantly cheaper than a re-factorization of the system at each new iterate. This paper focuses on the cheap update of a rectangular LU decomposition after a rank-1 modification.
Two different procedures for updating the LU factorization are presented in detail and compared regarding their costs of computation and their stability. Moreover we will introduce an extension of these algorithms which further improves the computation time. This turns out to be an excellent alternative to algorithms based on orthogonal transformations.
This paper concerns hyperbolic systems of two linear first-order PDEs in one space dimension with periodicity conditions in time and reflection boundary conditions in space.
The coefficients of the PDEs are supposed to be time independent, but allowed to be discontinuous with respect to the space variable. We construct two scales of Banach spaces (for the solutions and for the right hand sides of the equations, respectively) such that the problem can be modeled by means of Fredholm operators of index zero between corresponding spaces of the two scales.
The main tools of the proofs are separation of variables, integral representation of the solutions of the corresponding boundary value problems of the ODE systems and an abstract criterion for Fredholmness which seems to be new.
In this paper we discuss the numerical
solution of projected generalized Lyapunov
equations using the matrix sign function
method. Such equations arise in stability
analysis and control problems for
descriptor systems including model reduction
based on balanced truncation. It is known
that the matrix sign function method applied
to a matrix pencil $\lambda E-A$ converges
if and only if $\lambda E-A$ is of index at
most two. The convergence is quadratic if
$E$ is nonsingular, and it is linear,
otherwise. We will propose a modification
of the matrix sign function method that
converges quadratically for pencils of
arbitrary index. Numerical examples will be
presented to demonstrate the properties of
the modified method.
This paper is concerned with transparent boundary
conditions (TBCs) for the time-dependent Schrödinger equation
on a circular domain.
Discrete TBCs are introduced in the
numerical simulations of problems on unbounded domains in order to reduce
the computational domain to a finite region in order to make this problem feasible for numerical simulations.
The main focus of this article is on the
appropriate discretization of such
TBCs for the two-dimensional Schrödinger equation
in conjunction with a conservative Crank-Nicolson-type finite difference discretization.
The presented discrete TBCs yield an unconditionally stable
numerical scheme and are completely reflection-free at the boundary.
Furthermore we prove concisely the stability of the recurrence formulas used to
obtain the convolution coefficients of the new discrete TBC
for a spatially dependent potential.
In this paper we give an overview of model
order reduction techniques for coupled
systems. We consider linear time-invariant
control systems that are coupled through
input-output relations and discuss model
reduction of such systems using moment
matching and balanced truncation.
Structure-preserving approaches to model
order reduction of coupled systems are also
presented. Numerical examples are given.
Consistent Initialization and Perturbation Analysis for Abstract Differential-Algebraic Equations
(2006)
In this paper we consider linear and time-invariant
differential-algebraic equations (DAEs) $E\dot{x}(t)=Ax(t)+f(t)$,
$x(0)=x_0$, where $x(\cdot)$ and $f(\cdot)$ are functions with
values in separable Hilbert spaces $X$ and $Z$. $E:X\To Z$ is
assumed to be a bounded operator, whereas $A$ is closed and defined
on some dense subspace $D(A)$ which is in general a proper subset of
$X$. Based on a decoupling of the algebraic and the differential
part, the set of initial values being consistent with the given
inhomogeneity will be parameterized. As a consequence of these results, we
will derive estimates for the trajectory $x(\cdot)$ in dependence of the initial
state $x_0$ and the inhomogeneity $f(\cdot)$. In the theory of
differential-algebraic equations, this is commonly known as
perturbation analysis.
This work deals with the efficient numerical solution of
the two-dimensional one-way Helmholtz equation
posed on an unbounded domain.
In this case one has to introduce
artificial boundary conditions to confine the computational domain.
Here we construct with the Z-transformation
so-called discrete transparent
boundary conditions
for higher-order parabolic equations schemes.
These methods are Pade ``Parabolic'' approximations of
the one-way Helmholtz equation
and frequently used in integrated optics and (underwater) acoustics.
We present a domain decomposition approach for the computation of the
electromagnetic field within periodic structures. We use a
Schwarz method with transparent boundary conditions at the interfaces of
the domains. Transparent boundary conditions are approximated by the
perfectly matched layer method (PML). To cope with Wood anomalies
appearing in periodic structures an adaptive strategy to determine
optimal PML parameters is developed. \\ We focus on the application to
typical EUV lithography line masks. Light propagation within the
multi-layer stack of the EUV mask is treated analytically. This results
in a drastic reduction of the computational costs and allows for the
simulation of next generation lithography masks
on a standard personal computer.
In this work we derive an exact discrete artificial boundary condition
for the Crank-Nicolson scheme for solving the Black-Scholes
equation for the valuation of American options.
To ensure stability and to avoid any numerical reflections
we derive the artificial boundary
condition on a purely discrete level.
Since the exact discrete artificial boundary condition
includes a convolution with respect to time
with a weakly decaying kernel, its numerical evaluation
becomes very costly for large-time simulations.
As a remedy we construct approximate artificial boundary conditions
with a kernel having the form of a
finite sum-of-exponentials, which can be evaluated in a very
efficient recursion. We prove a simple stability criteria
for the approximated artificial boundary conditions.
Finally we illustrate the
efficiency of the proposed method on several examples
and compare it to previously obtained discretized artificial boundary conditions.