This thesis deals with the optimal control of PDEs. After a brief introduction in the theory of elliptic and parabolic PDEs, we introduce a software that solves systems of PDEs by the finite elements method. In the second chapter we derive optimality conditions in terms of function spaces, i.e. a systems of PDEs coupled by some pointwise relations. Now we present algorithms to solve the optimality systems numerically and present some numerical test cases. A further chapter deals with the so called lack of adjointness, an issue of gradient methods applied on parabolic optimal control problems. However, since optimal control problems lead to large numerical schemes, model reduction becomes popular. We analyze the proper orthogonal decomposition method and apply it to our model problems. Finally, we apply all considered techniques to a real world problem.

The work selects a specific issue from the numerical analysis of optimal control problems. We investigate a linear-quadratic optimal control problem based on a partial differential equation on 3-dimensional non-convex domains. Based on efficient solution methods for the partial differential equation an algorithm known from control theory is applied. Now the main objectives are to prove that there is no degradation in efficiency and to verify the result by numerical experiments. We describe a solution method which has second order convergence, although the intermediate control approximations are piecewise constant functions. This superconvergence property is gained from a special projection operator which generates a piecewise constant approximation that has a supercloseness property, from a sufficiently graded mesh which compensates the singularities introduced by the non-convex domain, and from a discretization condition which eliminates some pathological cases. Both isotropic and anisotropic discretizations are investigated and similar superconvergence properties are proven. A model problem is presented and important results from the regularity theory of solutions to partial differential equation in non-convex domains have been collected in the first chapters. Then a collection of statements from the finite element analysis and corresponding numerical solution strategies is given. Here we show newly developed tools regarding error estimates and projections into finite element spaces. These tools are necessary to achieve the main results. Known fundamental statements from control theory are applied to the given model problems and certain conditions on the discretization are defined. Then we describe the implementation used to solve the model problems and present all computed results.

We consider the solution of nonlinear optimal control problems subject to stochastic perturbations with incomplete observations. In particular, we generalize results obtained by Ito and Kunisch in [8] where they consider a receding horizon control (RHC) technique based on linearizing the problem on small intervals. The linear-quadratic optimal control problem for the resulting time-invariant (LTI) problem is then solved using the linear quadratic Gaussian (LQG) design. Here, we allow linearization about an instationary reference trajectory and thus obtain a linear time-varying (LTV) problem on each time horizon. Additionally, we apply a model predictive control (MPC) scheme which can be seen as a generalization of RHC and we allow covariance matrices of the noise processes not equal to the identity. We illustrate the MPC/LQG approach for a three dimensional reaction-diffusion system. In particular, we discuss the benefits of time-varying linearizations over time-invariant ones.

In this thesis, we derive necessary optimality conditions for bilevel programming problems (BPPs for short) in Banach spaces. This rather abstract setting reflects our desire to characterize the local optimal solutions of hierarchical optimization problems in function spaces arising from several applications. Since our considerations are based on the tools of variational analysis introduced by Boris Mordukhovich, we study related properties of pointwise defined sets in function spaces. The presence of sequential normal compactness for such sets in Lebesgue and Sobolev spaces as well as the variational geometry of decomposable sets in Lebesgue spaces is discussed. Afterwards, we investigate mathematical problems with complementarity constraints (MPCCs for short) in Banach spaces which are closely related to BPPs. We introduce reasonable stationarity concepts and constraint qualifications which can be used to handle MPCCs. The relations between the mentioned stationarity notions are studied in the setting where the underlying complementarity cone is polyhedric. The results are applied to the situations where the complementarity cone equals the nonnegative cone in a Lebesgue space or is polyhedral. Next, we use the three main approaches of transforming a BPP into a single-level program (namely the presence of a unique lower level solution, the KKT approach, and the optimal value approach) to derive necessary optimality conditions for BPPs. Furthermore, we comment on the relation between the original BPP and the respective surrogate problem. We apply our findings to formulate necessary optimality conditions for three different classes of BPPs. First, we study a BPP with semidefinite lower level problem possessing a unique solution. Afterwards, we deal with bilevel optimal control problems with dynamical systems of ordinary differential equations at both decision levels. Finally, an optimal control problem of ordinary or partial differential equations with implicitly given pointwise state constraints is investigated.

The work selects a specific issue from the numerical analysis of optimal control problems. We investigate a linear-quadratic optimal control problem based on a partial differential equation on 3-dimensional non-convex domains. Based on efficient solution methods for the partial differential equation an algorithm known from control theory is applied. Now the main objectives are to prove that there is no degradation in efficiency and to verify the result by numerical experiments. We describe a solution method which has second order convergence, although the intermediate control approximations are piecewise constant functions. This superconvergence property is gained from a special projection operator which generates a piecewise constant approximation that has a supercloseness property, from a sufficiently graded mesh which compensates the singularities introduced by the non-convex domain, and from a discretization condition which eliminates some pathological cases. Both isotropic and anisotropic discretizations are investigated and similar superconvergence properties are proven. A model problem is presented and important results from the regularity theory of solutions to partial differential equation in non-convex domains have been collected in the first chapters. Then a collection of statements from the finite element analysis and corresponding numerical solution strategies is given. Here we show newly developed tools regarding error estimates and projections into finite element spaces. These tools are necessary to achieve the main results. Known fundamental statements from control theory are applied to the given model problems and certain conditions on the discretization are defined. Then we describe the implementation used to solve the model problems and present all computed results.

We consider the solution of nonlinear optimal control problems subject to stochastic perturbations with incomplete observations. In particular, we generalize results obtained by Ito and Kunisch in [8] where they consider a receding horizon control (RHC) technique based on linearizing the problem on small intervals. The linear-quadratic optimal control problem for the resulting time-invariant (LTI) problem is then solved using the linear quadratic Gaussian (LQG) design. Here, we allow linearization about an instationary reference trajectory and thus obtain a linear time-varying (LTV) problem on each time horizon. Additionally, we apply a model predictive control (MPC) scheme which can be seen as a generalization of RHC and we allow covariance matrices of the noise processes not equal to the identity. We illustrate the MPC/LQG approach for a three dimensional reaction-diffusion system. In particular, we discuss the benefits of time-varying linearizations over time-invariant ones.

Wir betrachten Optimalsteuerungsprobleme, die von partiellen Differentialgleichungen beziehungsweise Variationsungleichungen mit Differentialoperatoren erster Ordnung abhängen. Wir führen die Reformulierung eines Tagebauplanungsproblems, das auf stetigen Funktionen beruht, ein. Das Resultat ist ein Optimalsteuerungsproblem für Viskositätslösungen einer Eikonalgleichung. Die Existenz von Lösungen dieses und bestimmter Hilfsprobleme, die von semilinearen PDG‘s mit künstlicher Viskosität abhängen, wird bewiesen, Stationaritätsbedingungen hergeleitet und ein schwaches Konsistenzresultat für stationäre Punkte präsentiert. Des Weiteren betrachten wir Optimalsteuerungsprobleme, die von stationären Variationsungleichungen erster Art mit linearen Differentialoperatoren erster Ordnung abhängen. Wir diskutieren Lösbarkeit und Stationaritätskonzepte für diese Probleme. Für letzteres vergleichen wir Ergebnisse, die entweder durch die Anwendung von Penalisierungs- und Regularisierungsansätzen direkt auf Ebene von Differentialoperatoren erster Ordnung oder als Grenzwertprozess von Stationaritätssystemen für viskositätsregularisierte Optimalsteuerungsprobleme unter passenden Annahmen erhalten werden. Um die Konsistenz von ursprünglichem und regularisierten Problemen zu sichern, wird ein bekanntes Ergebnis für Lösungen von VU’s mit degeneriertem Differentialoperator erweitert. In beiden Fällen ist die erhaltene Stationarität schwächer als W-stationarität. Die theoretischen Ergebnisse werden anhand numerischer Beispiele verifiziert. Wir erweitern diese Ergebnisse auf Optimalsteuerungsprobleme bezüglich zeitabhängiger VU’s mit Differentialoperatoren erster Ordnung. Hierfür wird die Existenz von Lösungen bewiesen und erneut ein Stationaritätssystem mit Hilfe verschwindender Viskositäten unter bestimmten Beschränktheitsannahmen hergeleitet. Die erhaltenen Ergebnisse werden anhand von numerischen Beispielen verifiziert. / We consider problems of optimal control subject to partial differential equations and variational inequality problems with first order differential operators. We introduce a reformulation of an open pit mine planning problem that is based on continuous functions. The resulting formulation is a problem of optimal control subject to viscosity solutions of a partial differential equation of Eikonal Type. The existence of solutions to this problem and auxiliary problems of optimal control subject to regularized, semilinear PDE’s with artificial viscosity is proven. For the latter a first order optimality condition is established and a mild consistency result for the stationary points is proven. Further we study certain problems of optimal control subject to time-independent variational inequalities of the first kind with linear first order differential operators. We discuss solvability and stationarity concepts for such problems. In the latter case, we compare the results obtained by either utilizing penalization-regularization strategies directly on the first order level or considering the limit of systems for viscosity-regularized problems under suitable assumptions. To guarantee the consistency of the original and viscosity-regularized problems of optimal control, we extend known results for solutions to variational inequalities with degenerated differential operators. In both cases, the resulting stationarity concepts are weaker than W-stationarity. We validate the theoretical findings by numerical experiments for several examples. Finally, we extend the results from the time-independent to the case of problems of optimal control subject to VI’s with linear first order differential operators that are time-dependent. After establishing the existence of solutions to the problem of optimal control, a stationarity system is derived by a vanishing viscosity approach under certain boundedness assumptions and the theoretical findings are validated by numerical experiments.

Die Arbeit abstrahiert bekannte Konzepte der endlichdimensionalen Optimierung im Hinblick auf deren Anwendung in Banachräumen. Hierfür werden zunächst grundlegende Elemente der Funktionalanalysis wie schwache Konvergenz, Dualräume und Reflexivität vorgestellt. Anschließend erfolgt eine kurze Einführung in die Thematik der Fréchet-Differenzierbarkeit und eine Abstraktion des Begriffs der partiellen Ordnungsrelation in normierten Räumen. Nach der Formulierung eines allgemeinen Existenzsatzes für globale Optimallösungen von abstrakten Optimierungsaufgaben werden notwendige Optimalitätsbedingungen vom Karush-Kuhn-Tucker-Typ hergeleitet. Abschließend wird eine hinreichende Optimalitätsbedingung vom Karush-Kuhn-Tucker-Typ unter verallgemeinerten Konvexitätsvoraussetzungen verifiziert.

Theoretical bounds for performance measures of thermodynamical systems are investigated under conditions of finite times and rates of processes using endoreversible models. These models consist of reversible operating sub-systems which exchange energy via generally irreversible interactions. Analytical and numerical calculations are performed to obtain performance optima and respective optimized process and design parameters for four model systems. A heat engine where the heat transfer between the working fluid and heat reservoirs is described by generalized, polytropic process is optimized for maximum work output. Thermal efficiencies, optimal values for temperatures and process times of the heat transfer processes are determined. A model of a generalized system suited to describe the operation of heat engines, refrigerators, and heat pumps is optimized with respect to thermal efficiency. Several examples illustrate how the results of the analysis are used to allocate financial resources to the heat exchanger inventory in an optimal way. A power-producing thermal system which exchanges heat with several heat reservoirs via irreversible heat transfer processes is analyzed to find the optimal contact times between the working fluid and each of the reservoirs. The piston motion of a Diesel engine is optimized to achieve maximum work for a given amount of fuel. The endoreversible model of the Diesel engine accounts for the temporal variations of the heat produced by the combustion process, the basic flow pattern within the engine's cylinder, the temperature dependence of the viscosity, thermal conductivity, and heat capacity of the working fluid and losses due to friction and heat leak through the cylinder walls. / Theoretische Grenzen für verschiedene Leistungsmerkmale von thermodynamischen Systemen werden unter der Bedingung endlicher Zeiten und Prozessraten im Rahmen endoreversibler Modelle untersucht. Diese Modelle bestehen aus reversiblen Subsystemen, welche über allgemein irreversible Wechselwirkungen Energie austauschen. Analytische und nummerische Berechnungen quantifizieren diese Grenzen und liefern optimale Prozess- und Konstruktionsparameter für vier Modellsysteme. Für eine auf maximale Ausgangsarbeit optimierte Wärmekraftmaschine, bei der die Wärme zwischen Arbeitsmedium und Wärmereservoirs während allgemeiner polytroper Zustandsänderungen des Arbeitsmediums übertragen wird, werden optimale Temperaturen und Zeiten für die Wärmeübertragungsprozesse sowie die thermischen Wirkungsgrade bestimmt. Für ein wirkungsgrad-optimiertes Modell eines verallgemeinerten thermischen Umwandlungssytems, das sowohl Wärmekraftmaschinen, Kühler und Wärmepumpen beschreibt, wird die optimale Verteilung von Investitionskosten auf die Wärmetauscher ermittelt und die Anwendung der allgemeingültigen Ergebnisse anhand mehrerer Beispiele demonstriert. Für eine Wärmekraftmaschine mit mehreren Wärmereservoirs wird bestimmt, welche der Wärmereservoirs wie lange kontaktiert werden müssen, um eine maximale Ausgangsarbeit zu erzielen. Für einen Dieselmotor wird die Kolbenbewegung so optimiert, dass bei gegebener Treibstoffmenge eine maximale Ausgangsarbeit erzielt wird. Das endoreversible Modell des Dieselmotors berücksichtigt die Temperaturabhängigkeit der Wärmekapazität, Wärmeleitfähigkeit und Viskosität des Arbeitsfluids, die Zeitabhängigkeit des Verbrennungsprozesses sowie Reibungs- und Wärmeverluste.

It is well-known that the solution of a so-called singularly perturbed differential equation exhibits layers. These are small regions in the domain where the solution changes drastically. These layers deteriorate the convergence of standard numerical algorithms, such as the finite element method on a uniform mesh. In the past many approaches were developed to overcome this difficulty. In this context it was very helpful to understand the structure of the solution - especially to know where the layers can occur. Therefore, we have a lot of analysis in the literature concerning the properties of solutions of such problems. Nevertheless, this field is far from being understood conclusively. More recently, there is an increasing interest in the numerics of optimal control problems subject to a singularly perturbed convection-diffusion equation and box constraints for the control. However, it is not much known about the solutions of such optimal control problems. The proposed solution methods are based on the experience one has from scalar singularly perturbed differential equations, but so far, the analysis presented does not use the structure of the solution and in fact, the provided bounds are rather meaningless for solutions which exhibit boundary layers, since these bounds scale like epsilon^(-1.5) as epsilon converges to 0. In this thesis we strive to prove bounds for the solution and its derivatives of the optimal control problem. These bounds show that there is an additional layer that is weaker than the layers one expects knowing the results for scalar differential equation problems, but that weak layer deteriorates the convergence of the proposed methods. In Chapter 1 and 2 we discuss the optimal control problem for the one-dimensional case. We consider the case without control constraints and the case with control constraints separately. For the case without control constraints we develop a method to prove bounds for arbitrary derivatives of the solution, given the data is smooth enough. For the latter case we prove bounds for the derivatives up to the second order. Subsequently, we discuss several discretization methods. In this context we use special Shishkin meshes. These meshes are piecewise equidistant, but have a very fine subdivision in the region of the layers. Additionally, we consider different ways of discretizing the control constraints. The first one enforces the compliance of the constraints everywhere and the other one enforces it only in the mesh nodes. For each proposed algorithm we prove convergence estimates that are independent of the parameter epsilon. Hence, they are meaningful even for small values of epsilon. As a next step we turn to the two-dimensional case. To be able to adapt the proofs of Chapter 2 to this case we require bounds for the solution of the scalar differential equation problem for a right hand side f only in W^(1,infty). Although, a lot of results for this problem can be found in the literature but we can not apply any of them, because they require a smooth right hand side f in C^(2,alpha) for some alpha in (0,1). Therefore, we dedicate Chapter 3 to the analysis of the scalar differential equations problem only using a right hand side f that is not very smooth. In Chapter 4 we strive to prove bounds for the solution of the optimal control problem in the two dimensional case. The analysis for this problem is not complete. Especially, the characteristic layers induce subproblems that are not understood completely. Hence, we can not prove sharp bounds for all terms in the solution decomposition we construct. Nevertheless, we propose a solution method. Numerical results indicate an epsilon-independent convergence for the considered examples - although we are not able to prove this.