510 Mathematik
Refine
Year of publication
Document Type
- Doctoral Thesis (77)
- Habilitation (2)
- Article (1)
Keywords
- Optimierung (13)
- Approximationstheorie (7)
- Approximation (6)
- Funktionentheorie (6)
- Partielle Differentialgleichung (6)
- Universalität (6)
- Funktionalanalysis (5)
- universal functions (5)
- Analysis (4)
- Numerische Mathematik (4)
Institute
- Mathematik (63)
- Fachbereich 4 (16)
Traditionell werden Zufallsstichprobenerhebungen so geplant, dass nationale Statistiken zuverlässig mit einer adäquaten Präzision geschätzt werden können. Hierbei kommen vorrangig designbasierte, Modell-unterstützte (engl. model assisted) Schätzmethoden zur Anwendung, die überwiegend auf asymptotischen Eigenschaften beruhen. Für kleinere Stichprobenumfänge, wie man sie für Small Areas (Domains bzw. Subpopulationen) antrifft, eignen sich diese Schätzmethoden eher nicht, weswegen für diese Anwendung spezielle modellbasierte Small Area-Schätzverfahren entwickelt wurden. Letztere können zwar Verzerrungen aufweisen, besitzen jedoch häufig einen kleineren mittleren quadratischen Fehler der Schätzung als dies für designbasierte Schätzer der Fall ist. Den Modell-unterstützten und modellbasierten Methoden ist gemeinsam, dass sie auf statistischen Modellen beruhen; allerdings in unterschiedlichem Ausmass. Modell-unterstützte Verfahren sind in der Regel so konstruiert, dass der Beitrag des Modells bei sehr grossen Stichprobenumfängen gering ist (bei einer Grenzwertbetrachtung sogar wegfällt). Bei modellbasierten Methoden nimmt das Modell immer eine tragende Rolle ein, unabhängig vom Stichprobenumfang. Diese Überlegungen veranschaulichen, dass das unterstellte Modell, präziser formuliert, die Güte der Modellierung für die Qualität der Small Area-Statistik von massgeblicher Bedeutung ist. Wenn es nicht gelingt, die empirischen Daten durch ein passendes Modell zu beschreiben und mit den entsprechenden Methoden zu schätzen, dann können massive Verzerrungen und / oder ineffiziente Schätzungen resultieren.
Die vorliegende Arbeit beschäftigt sich mit der zentralen Frage der Robustheit von Small Area-Schätzverfahren. Als robust werden statistische Methoden dann bezeichnet, wenn sie eine beschränkte Einflussfunktion und einen möglichst hohen Bruchpunkt haben. Vereinfacht gesprochen zeichnen sich robuste Verfahren dadurch aus, dass sie nur unwesentlich durch Ausreisser und andere Anomalien in den Daten beeinflusst werden. Die Untersuchung zur Robustheit konzentriert sich auf die folgenden Modelle bzw. Schätzmethoden:
i) modellbasierte Schätzer für das Fay-Herriot-Modell (Fay und Herrot, 1979, J. Amer. Statist. Assoc.) und das elementare Unit-Level-Modell (vgl. Battese et al., 1988, J. Amer. Statist. Assoc.).
ii) direkte, Modell-unterstützte Schätzer unter der Annahme eines linearen Regressionsmodells.
Das Unit-Level-Modell zur Mittelwertschätzung beruht auf einem linearen gemischten Gauss'schen Modell (engl. mixed linear model, MLM) mit blockdiagonaler Kovarianzmatrix. Im Gegensatz zu bspw. einem multiplen linearen Regressionsmodell, besitzen MLM-Modelle keine nennenswerten Invarianzeigenschaften, so dass eine Kontamination der abhängigen Variablen unvermeidbar zu verzerrten Parameterschätzungen führt. Für die Maximum-Likelihood-Methode kann die resultierende Verzerrung nahezu beliebig groß werden. Aus diesem Grund haben Richardson und Welsh (1995, Biometrics) die robusten Schätzmethoden RML 1 und RML 2 entwickelt, die bei kontaminierten Daten nur eine geringe Verzerrung aufweisen und wesentlich effizienter sind als die Maximum-Likelihood-Methode. Eine Abwandlung von Methode RML 2 wurde Sinha und Rao (2009, Canad. J. Statist.) für die robuste Schätzung von Unit-Level-Modellen vorgeschlagen. Allerdings erweisen sich die gebräuchlichen numerischen Verfahren zur Berechnung der RML-2-Methode (dies gilt auch für den Vorschlag von Sinha und Rao) als notorisch unzuverlässig. In dieser Arbeit werden zuerst die Konvergenzprobleme der bestehenden Verfahren erörtert und anschließend ein numerisches Verfahren vorgeschlagen, das sich durch wesentlich bessere numerische Eigenschaften auszeichnet. Schließlich wird das vorgeschlagene Schätzverfahren im Rahmen einer Simulationsstudie untersucht und anhand eines empirischen Beispiels zur Schätzung von oberirdischer Biomasse in norwegischen Kommunen illustriert.
Das Modell von Fay-Herriot kann als Spezialfall eines MLM mit blockdiagonaler Kovarianzmatrix aufgefasst werden, obwohl die Varianzen des Zufallseffekts für die Small Areas nicht geschätzt werden müssen, sondern als bereits bekannte Größen betrachtet werden. Diese Eigenschaft kann man sich nun zunutze machen, um die von Sinha und Rao (2009) vorgeschlagene Robustifizierung des Unit-Level-Modells direkt auf das Fay-Herriot Model zu übertragen. In der vorliegenden Arbeit wird jedoch ein alternativer Vorschlag erarbeitet, der von der folgenden Beobachtung ausgeht: Fay und Herriot (1979) haben ihr Modell als Verallgemeinerung des James-Stein-Schätzers motiviert, wobei sie sich einen empirischen Bayes-Ansatz zunutze machen. Wir greifen diese Motivation des Problems auf und formulieren ein analoges robustes Bayes'sches Verfahren. Wählt man nun in der robusten Bayes'schen Problemformulierung die ungünstigste Verteilung (engl. least favorable distribution) von Huber (1964, Ann. Math. Statist.) als A-priori-Verteilung für die Lokationswerte der Small Areas, dann resultiert als Bayes-Schätzer [=Schätzer mit dem kleinsten Bayes-Risk] die Limited-Translation-Rule (LTR) von Efron und Morris (1971, J. Amer. Statist. Assoc.). Im Kontext der frequentistischen Statistik kann die Limited-Translation-Rule nicht verwendet werden, weil sie (als Bayes-Schätzer) auf unbekannten Parametern beruht. Die unbekannten Parameter können jedoch nach dem empirischen Bayes-Ansatz an der Randverteilung der abhängigen Variablen geschätzt werden. Hierbei gilt es zu beachten (und dies wurde in der Literatur vernachlässigt), dass die Randverteilung unter der ungünstigsten A-priori-Verteilung nicht einer Normalverteilung entspricht, sondern durch die ungünstigste Verteilung nach Huber (1964) beschrieben wird. Es ist nun nicht weiter erstaunlich, dass es sich bei den Maximum-Likelihood-Schätzern von Regressionskoeffizienten und Modellvarianz unter der Randverteilung um M-Schätzer mit der Huber'schen psi-Funktion handelt.
Unsere theoriegeleitete Herleitung von robusten Schätzern zum Fay-Herriot-Modell zeigt auf, dass bei kontaminierten Daten die geschätzte LTR (mit Parameterschätzungen nach der M-Schätzmethodik) optimal ist und, dass die LTR ein integraler Bestandteil der Schätzmethodik ist (und nicht als ``Zusatz'' o.Ä. zu betrachten ist, wie dies andernorts getan wird). Die vorgeschlagenen M-Schätzer sind robust bei Vorliegen von atypischen Small Areas (Ausreissern), wie dies auch die Simulations- und Fallstudien zeigen. Um auch Robustheit bei Vorkommen von einflussreichen Beobachtungen in den unabhängigen Variablen zu erzielen, wurden verallgemeinerte M-Schätzer (engl. generalized M-estimator) für das Fay-Herriot-Modell entwickelt.
Many NP-hard optimization problems that originate from classical graph theory, such as the maximum stable set problem and the maximum clique problem, have been extensively studied over the past decades and involve the choice of a subset of edges or vertices. There usually exist combinatorial methods that can be applied to solve them directly in the graph.
The most simple method is to enumerate feasible solutions and select the best. It is not surprising that this method is very slow oftentimes, so the task is to cleverly discard fruitless search space during the search. An alternative method to solve graph problems is to formulate integer linear programs, such that their solution yields an optimal solution to the original optimization problem in the graph. In order to solve integer linear programs, one can start with relaxing the integer constraints and then try to find inequalities for cutting off fractional extreme points. In the best case, it would be possible to describe the convex hull of the feasible region of the integer linear program with a set of inequalities. In general, giving a complete description of this convex hull is out of reach, even if it has a polynomial number of facets. Thus, one tries to strengthen the (weak) relaxation of the integer linear program best possible via strong inequalities that are valid for the convex hull of feasible integer points.
Many classes of valid inequalities are of exponential size. For instance, a graph can have exponentially many odd cycles in general and therefore the number of odd cycle inequalities for the maximum stable set problem is exponential. It is sometimes possible to check in polynomial time if some given point violates any of the exponentially many inequalities. This is indeed the case for the odd cycle inequalities for the maximum stable set problem. If a polynomial time separation algorithm is known, there exists a formulation of polynomial size that contains a given point if and only if it does not violate one of the (potentially exponentially many) inequalities. This thesis can be divided into two parts. The first part is the main part and it contains various new results. We present new extended formulations for several optimization problems, i.e. the maximum stable set problem, the nonconvex quadratic program with box
constraints and the p-median problem. In the second part we modify a very fast algorithm for finding a maximum clique in very large sparse graphs. We suggest and compare three alternative versions of this algorithm to the original version and compare their strengths and weaknesses.
Data used for the purpose of machine learning are often erroneous. In this thesis, p-quasinorms (p<1) are employed as loss functions in order to increase the robustness of training algorithms for artificial neural networks. Numerical issues arising from these loss functions are addressed via enhanced optimization algorithms (proximal point methods; Frank-Wolfe methods) based on the (non-monotonic) Armijo-rule. Numerical experiments comprising 1100 test problems confirm the effectiveness of the approach. Depending on the parametrization, an average reduction of the absolute residuals of up to 64.6% is achieved (aggregated over 100 test problems).
In this thesis we study structure-preserving model reduction methods for the efficient and reliable approximation of dynamical systems. A major focus is the approximation of a nonlinear flow problem on networks, which can, e.g., be used to describe gas network systems. Our proposed approximation framework guarantees so-called port-Hamiltonian structure and is general enough to be realizable by projection-based model order reduction combined with complexity reduction. We divide the discussion of the flow problem into two parts, one concerned with the linear damped wave equation and the other one with the general nonlinear flow problem on networks.
The study around the linear damped wave equation relies on a Galerkin framework, which allows for convenient network generalizations. Notable contributions of this part are the profound analysis of the algebraic setting after space-discretization in relation to the infinite dimensional setting and its implications for model reduction. In particular, this includes the discussion of differential-algebraic structures associated to the network-character of our problem and the derivation of compatibility conditions related to fundamental physical properties. Amongst the different model reduction techniques, we consider the moment matching method to be a particularly well-suited choice in our framework.
The Galerkin framework is then appropriately extended to our general nonlinear flow problem. Crucial supplementary concepts are required for the analysis, such as the partial Legendre transform and a more careful discussion of the underlying energy-based modeling. The preservation of the port-Hamiltonian structure after the model-order- and complexity-reduction-step represents a major focus of this work. Similar as in the analysis of the model order reduction, compatibility conditions play a crucial role in the analysis of our complexity reduction, which relies on a quadrature-type ansatz. Furthermore, energy-stable time-discretization schemes are derived for our port-Hamiltonian approximations, as structure-preserving methods from literature are not applicable due to our rather unconventional parametrization of the solution.
Apart from the port-Hamiltonian approximation of the flow problem, another topic of this thesis is the derivation of a new extension of moment matching methods from linear systems to quadratic-bilinear systems. Most system-theoretic reduction methods for nonlinear systems rely on multivariate frequency representations. Our approach instead uses univariate frequency representations tailored towards user-defined families of inputs. Then moment matching corresponds to a one-dimensional interpolation problem rather than to a multi-dimensional interpolation as for the multivariate approaches, i.e., it involves fewer interpolation frequencies to be chosen. The notion of signal-generator-driven systems, variational expansions of the resulting autonomous systems as well as the derivation of convenient tensor-structured approximation conditions are the main ingredients of this part. Notably, our approach allows for the incorporation of general input relations in the state equations, not only affine-linear ones as in existing system-theoretic methods.
Nonlocal operators are used in a wide variety of models and applications due to many natural phenomena being driven by nonlocal dynamics. Nonlocal operators are integral operators allowing for interactions between two distinct points in space. The nonlocal models investigated in this thesis involve kernels that are assumed to have a finite range of nonlocal interactions. Kernels of this type are used in nonlocal elasticity and convection-diffusion models as well as finance and image analysis. Also within the mathematical theory they arouse great interest, as they are asymptotically related to fractional and classical differential equations.
The results in this thesis can be grouped according to the following three aspects: modeling and analysis, discretization and optimization.
Mathematical models demonstrate their true usefulness when put into numerical practice. For computational purposes, it is important that the support of the kernel is clearly determined. Therefore nonlocal interactions are typically assumed to occur within an Euclidean ball of finite radius. In this thesis we consider more general interaction sets including norm induced balls as special cases and extend established results about well-posedness and asymptotic limits.
The discretization of integral equations is a challenging endeavor. Especially kernels which are truncated by Euclidean balls require carefully designed quadrature rules for the implementation of efficient finite element codes. In this thesis we investigate the computational benefits of polyhedral interaction sets as well as geometrically approximated interaction sets. In addition to that we outline the computational advantages of sufficiently structured problem settings.
Shape optimization methods have been proven useful for identifying interfaces in models governed by partial differential equations. Here we consider a class of shape optimization problems constrained by nonlocal equations which involve interface-dependent kernels. We derive the shape derivative associated to the nonlocal system model and solve the problem by established numerical techniques.
We consider a linear regression model for which we assume that some of the observed variables are irrelevant for the prediction. Including the wrong variables in the statistical model can either lead to the problem of having too little information to properly estimate the statistic of interest, or having too much information and consequently describing fictitious connections. This thesis considers discrete optimization to conduct a variable selection. In light of this, the subset selection regression method is analyzed. The approach gained a lot of interest in recent years due to its promising predictive performance. A major challenge associated with the subset selection regression is the computational difficulty. In this thesis, we propose several improvements for the efficiency of the method. Novel bounds on the coefficients of the subset selection regression are developed, which help to tighten the relaxation of the associated mixed-integer program, which relies on a Big-M formulation. Moreover, a novel mixed-integer linear formulation for the subset selection regression based on a bilevel optimization reformulation is proposed. Finally, it is shown that the perspective formulation of the subset selection regression is equivalent to a state-of-the-art binary formulation. We use this insight to develop novel bounds for the subset selection regression problem, which show to be highly effective in combination with the proposed linear formulation.
In the second part of this thesis, we examine the statistical conception of the subset selection regression and conclude that it is misaligned with its intention. The subset selection regression uses the training error to decide on which variables to select. The approach conducts the validation on the training data, which oftentimes is not a good estimate of the prediction error. Hence, it requires a predetermined cardinality bound. Instead, we propose to select variables with respect to the cross-validation value. The process is formulated as a mixed-integer program with the sparsity becoming subject of the optimization. Usually, a cross-validation is used to select the best model out of a few options. With the proposed program the best model out of all possible models is selected. Since the cross-validation is a much better estimate of the prediction error, the model can select the best sparsity itself.
The thesis is concluded with an extensive simulation study which provides evidence that discrete optimization can be used to produce highly valuable predictive models with the cross-validation subset selection regression almost always producing the best results.
In this thesis, we consider the solution of high-dimensional optimization problems with an underlying low-rank tensor structure. Due to the exponentially increasing computational complexity in the number of dimensions—the so-called curse of dimensionality—they present a considerable computational challenge and become infeasible even for moderate problem sizes.
Multilinear algebra and tensor numerical methods have a wide range of applications in the fields of data science and scientific computing. Due to the typically large problem sizes in practical settings, efficient methods, which exploit low-rank structures, are essential. In this thesis, we consider an application each in both of these fields.
Tensor completion, or imputation of unknown values in partially known multiway data is an important problem, which appears in statistics, mathematical imaging science and data science. Under the assumption of redundancy in the underlying data, this is a well-defined problem and methods of mathematical optimization can be applied to it.
Due to the fact that tensors of fixed rank form a Riemannian submanifold of the ambient high-dimensional tensor space, Riemannian optimization is a natural framework for these problems, which is both mathematically rigorous and computationally efficient.
We present a novel Riemannian trust-region scheme, which compares favourably with the state of the art on selected application cases and outperforms known methods on some test problems.
Optimization problems governed by partial differential equations form an area of scientific computing which has applications in a variety of areas, ranging from physics to financial mathematics. Due to the inherent high dimensionality of optimization problems arising from discretized differential equations, these problems present computational challenges, especially in the case of three or more dimensions. An even more challenging class of optimization problems has operators of integral instead of differential type in the constraint. These operators are nonlocal, and therefore lead to large, dense discrete systems of equations. We present a novel solution method, based on separation of spatial dimensions and provably low-rank approximation of the nonlocal operator. Our approach allows the solution of multidimensional problems with a complexity which is only slightly larger than linear in the univariate grid size; this improves the state of the art for a particular test problem problem by at least two orders of magnitude.
Many combinatorial optimization problems on finite graphs can be formulated as conic convex programs, e.g. the stable set problem, the maximum clique problem or the maximum cut problem. Especially NP-hard problems can be written as copositive programs. In this case the complexity is moved entirely into the copositivity constraint.
Copositive programming is a quite new topic in optimization. It deals with optimization over the so-called copositive cone, a superset of the positive semidefinite cone, where the quadratic form x^T Ax has to be nonnegative for only the nonnegative vectors x. Its dual cone is the cone of completely positive matrices, which includes all matrices that can be decomposed as a sum of nonnegative symmetric vector-vector-products.
The related optimization problems are linear programs with matrix variables and cone constraints.
However, some optimization problems can be formulated as combinatorial problems on infinite graphs. For example, the kissing number problem can be formulated as a stable set problem on a circle.
In this thesis we will discuss how the theory of copositive optimization can be lifted up to infinite dimension. For some special cases we will give applications in combinatorial optimization.
Optimal Control of Partial Integro-Differential Equations and Analysis of the Gaussian Kernel
(2018)
An important field of applied mathematics is the simulation of complex financial, mechanical, chemical, physical or medical processes with mathematical models. In addition to the pure modeling of the processes, the simultaneous optimization of an objective function by changing the model parameters is often the actual goal. Models in fields such as finance, biology or medicine benefit from this optimization step.
While many processes can be modeled using an ordinary differential equation (ODE), partial differential equations (PDEs) are needed to optimize heat conduction and flow characteristics, spreading of tumor cells in tissue as well as option prices. A partial integro-differential equation (PIDE) is a parital differential equation involving an integral operator, e.g., the convolution of the unknown function with a given kernel function. PIDEs occur for example in models that simulate adhesive forces between cells or option prices with jumps.
In each of the two parts of this thesis, a certain PIDE is the main object of interest. In the first part, we study a semilinear PIDE-constrained optimal control problem with the aim to derive necessary optimality conditions. In the second, we analyze a linear PIDE that includes the convolution of the unknown function with the Gaussian kernel.
The economic growth theory analyses which factors affect economic growth and tries to analyze how it can last. A popular neoclassical growth model is the Ramsey-Cass-Koopmans model, which aims to determine how much of its income a nation or an economy should save in order to maximize its welfare. In this thesis, we present and analyze an extended capital accumulation equation of a spatial version of the Ramsey model, balancing diffusive and agglomerative effects. We model the capital mobility in space via a nonlocal diffusion operator which allows for jumps of the capital stock from one location to an other. Moreover, this operator smooths out heterogeneities in the factor distributions slower, which generated a more realistic behavior of capital flows. In addition to that, we introduce an endogenous productivity-production operator which depends on time and on the capital distribution in space. This operator models the technological progress of the economy. The resulting mathematical model is an optimal control problem under a semilinear parabolic integro-differential equation with initial and volume constraints, which are a nonlocal analog to local boundary conditions, and box-constraints on the state and the control variables. In this thesis, we consider this problem on a bounded and unbounded spatial domain, in both cases with a finite time horizon. We derive existence results of weak solutions for the capital accumulation equations in both settings and we proof the existence of a Ramsey equilibrium in the unbounded case. Moreover, we solve the optimal control problem numerically and discuss the results in the economic context.
