Data integration has been, and still is, a challenge for applications processing multiple heterogeneous data sources. Across the domains of schemas, ontologies, and metamodels, this imposes the need for mapping specifications, i.e. the task of discovering semantic correspondences between elements. Support for the development of such mappings has been researched, producing matching systems that automatically propose mapping suggestions. However, especially in the context of metamodel matching the result quality of state of the art matching techniques leaves room for improvement. Although the traditional approach of pair-wise element comparison works on smaller data sets, its quadratic complexity leads to poor runtime and memory performance and eventually to the inability to match, when applied on real-world data. The work presented in this thesis seeks to address these shortcomings. Thereby, we take advantage of the graph structure of metamodels. Consequently, we derive a planar graph edit distance as metamodel similarity metric and mining-based matching to make use of redundant information. We also propose a planar graph-based partitioning to cope with large-scale matching. These techniques are then evaluated using real-world mappings from SAP business integration scenarios and the MDA community. The results demonstrate improvement in quality and managed runtime and memory consumption for large-scale metamodel matching.

No description available.

Auf politischen Landkarten werden für gewöhnlich aneinandergrenzende Länder mit unterschiedlichen Farben gefärbt. Allerdings haben durchaus verschiedene Länder die gleiche Farbe, wenn sie keine gemeinsame Grenze haben. Doch wie viele verschiedene Farben benötigt man mindestens, um eine Karte nach diesen Regeln zu färben? Die beiden Mathematiker Appel und Haken haben gezeigt, dass für jede Landkarte in der Ebene immer vier Farben ausreichen. Dieser Beweis kommt allerdings nicht ohne intensiven Computereinsatz aus.

Werden in einem Graphen Knoten entfernt, so müssen auch alle Kanten entfernt werden, die diesen Knoten beinhalten. Dies kann dazu führen, dass Graphen nicht mehr zusammenhängend sind oder sich die Pfadlänge zwischen zwei Knoten verlängert. Um diesen Problemen entgegen zu wirken, müssen entsprechend der Graphenstruktur neue Kanten gezogen werden. Im Rahmen dieser Bachelorarbeit wurde ein Algorithmus entwickelt, der diese Kanten nach festen Regeln erstellt und somit die Struktur eines Graphen erhält, auch wenn Knoten aus diesem entfernt werden.

Bei der Partitionierung von Graphen versucht man, Strukturen in Graphen zu finden (etwa 3-Färbungen oder kleine Bisektionen). Mithilfe von Eigenwerten und Eigenvektoren können solche Probleme oftmals effizient gelöst werden. Wir stellen einen Algorithmus vor, der auf einem sehr allgemeinen Modell für zufällige Graphen bewiesenermaßen sehr gute Dienste leistet. Weiterhin untersuchen wir zufällige 3Sat-Formeln. Hier wollen wir mit Eigenwerten obere Schranken an die Anzahl der erfüllbaren Klauseln finden. Die gefundenen Schranken sind (in den meisten Fällen) nahezu optimal.

Graphene, the two-dimensional lattice of sp2-hybridized carbon atoms, has a great potential for future electronics, in particular for opto-electronic devices. The carrier relaxation dynamics, which is of key importance for such applications, is in the main focus of this thesis. Besides a short introduction into the most prominent material properties of graphene and the experimental techniques, this thesis is divided into three main parts. The investigation of the carrier relaxation dynamics in the absence of a magnetic field is presented in Chapter 3. In the first experiment, the anisotropy of the carrier excitation and relaxation in momentum space was investigated by pump-probe measurements in the near-infrared range. While this anisotropy was not considered in all previous experiments, our measurements with a temporal resolution of less than 50 fs revealed the polarization dependence of the carrier excitation and the subsequent relaxation. About 150 fs after the electrons are excited, the carrier distribution in momentum space gets isotropic, caused by electron-phonon scattering. In a second set of two-color pump-probe experiments, the temperature of the hot carrier distribution, which was obtained within the duration of the pump pulse (about 200 fs), could be estimated. Furthermore, a change in sign of the pump-probe signal can be used as an indicator for the Fermi energy of different graphene layers. Pump-probe experiments in the far-infrared range in reflection and transmission geometry were performed at high pump power. A strong saturation of the pump-induced transmission was found in previous experiments, which was attributed to the pump-induced change in absorption. Our investigation shows the strong influence of pump-induced reflection at long wavelengths, as well as a lot smaller influence of the saturation of the pump-induced change in absorption. At a high pump power, the increase of the reflection exceeds the change in absorption strongly, which leads to negative pump-probe signals in transmission geometry. In Chapter 4, investigations of the carrier dynamics of graphene in magnetic fields of up to 7T are presented. Even though the optical properties of Landau-quantized graphene are very interesting, the carrier dynamics were nearly unexplored. A low photon energy of 14meV allows the investigation of the intraband Landau-level (LL) transitions. These experiments revealed two main findings: Firstly, the Landau quantization strongly suppresses the carrier relaxation via optical-phonon scattering, resulting in an increased relaxation time. Secondly, a change in sign of the pump-probe signal can be observed when the magnetic field is varied. This change in sign indicates a hot carrier distribution shortly after the pump pulse, which means that carrier-carrier scattering remains very strong in magnetic fields. In a second set of pump-probe measurements, carried out at a photon energy of 75meV, the relaxation dynamics of interband LL transitions was investigated. In particular, experiments on the two energetically degenerate LL transitions LL(−1)->LL(0) and LL(0)->LL(1) showed the influence of extremely strong Auger processes. An ultrafast and extremely broadband terahertz detector, based on a graphene flake, is presented in the last chapter of this thesis. To couple the radiation efficiently to the small flake, the inner part of a logarithmic periodic antenna is connected to it. With a rise time of about 50 ps in a wavelength range of 9 μm to 500 μm, this detector is very interesting to obtain the temporal overlap in two-color pump-probe experiments with the free-electron laser FELBE. Furthermore, the importance of the substrate material, in particular for the high-speed performance, is discussed.

Graphene, the two-dimensional lattice of sp2-hybridized carbon atoms, has a great potential for future electronics, in particular for opto-electronic devices. The carrier relaxation dynamics, which is of key importance for such applications, is in the main focus of this thesis. Besides a short introduction into the most prominent material properties of graphene and the experimental techniques, this thesis is divided into three main parts. The investigation of the carrier relaxation dynamics in the absence of a magnetic field is presented in Chapter 3. In the first experiment, the anisotropy of the carrier excitation and relaxation in momentum space was investigated by pump-probe measurements in the near-infrared range. While this anisotropy was not considered in all previous experiments, our measurements with a temporal resolution of less than 50 fs revealed the polarization dependence of the carrier excitation and the subsequent relaxation. About 150 fs after the electrons are excited, the carrier distribution in momentum space gets isotropic, caused by electron-phonon scattering. In a second set of two-color pump-probe experiments, the temperature of the hot carrier distribution, which was obtained within the duration of the pump pulse (about 200 fs), could be estimated. Furthermore, a change in sign of the pump-probe signal can be used as an indicator for the Fermi energy of different graphene layers. Pump-probe experiments in the far-infrared range in reflection and transmission geometry were performed at high pump power. A strong saturation of the pump-induced transmission was found in previous experiments, which was attributed to the pump-induced change in absorption. Our investigation shows the strong influence of pump-induced reflection at long wavelengths, as well as a lot smaller influence of the saturation of the pump-induced change in absorption. At a high pump power, the increase of the reflection exceeds the change in absorption strongly, which leads to negative pump-probe signals in transmission geometry. In Chapter 4, investigations of the carrier dynamics of graphene in magnetic fields of up to 7T are presented. Even though the optical properties of Landau-quantized graphene are very interesting, the carrier dynamics were nearly unexplored. A low photon energy of 14meV allows the investigation of the intraband Landau-level (LL) transitions. These experiments revealed two main findings: Firstly, the Landau quantization strongly suppresses the carrier relaxation via optical-phonon scattering, resulting in an increased relaxation time. Secondly, a change in sign of the pump-probe signal can be observed when the magnetic field is varied. This change in sign indicates a hot carrier distribution shortly after the pump pulse, which means that carrier-carrier scattering remains very strong in magnetic fields. In a second set of pump-probe measurements, carried out at a photon energy of 75meV, the relaxation dynamics of interband LL transitions was investigated. In particular, experiments on the two energetically degenerate LL transitions LL(−1)->LL(0) and LL(0)->LL(1) showed the influence of extremely strong Auger processes. An ultrafast and extremely broadband terahertz detector, based on a graphene flake, is presented in the last chapter of this thesis. To couple the radiation efficiently to the small flake, the inner part of a logarithmic periodic antenna is connected to it. With a rise time of about 50 ps in a wavelength range of 9 μm to 500 μm, this detector is very interesting to obtain the temporal overlap in two-color pump-probe experiments with the free-electron laser FELBE. Furthermore, the importance of the substrate material, in particular for the high-speed performance, is discussed.

Graph polynomials are polynomials associated to graphs that encode the number of subgraphs with given properties. We list different frameworks used to define graph polynomials in the literature. We present the edge elimination polynomial and introduce several graph polynomials equivalent to it. Thereby, we connect a recursive definition to the counting of colorings and to the counting of (spanning) subgraphs. Furthermore, we define a graph polynomial that not only generalizes the mentioned, but also many of the well-known graph polynomials, including the Potts model, the matching polynomial, the trivariate chromatic polynomial and the subgraph component polynomial. We proof a recurrence relation for this graph polynomial using edge and vertex operation. The definitions and statements are given in such a way that most of them are also valid in the case of hypergraphs.

Data integration has been, and still is, a challenge for applications processing multiple heterogeneous data sources. Across the domains of schemas, ontologies, and metamodels, this imposes the need for mapping specifications, i.e. the task of discovering semantic correspondences between elements. Support for the development of such mappings has been researched, producing matching systems that automatically propose mapping suggestions. However, especially in the context of metamodel matching the result quality of state of the art matching techniques leaves room for improvement. Although the traditional approach of pair-wise element comparison works on smaller data sets, its quadratic complexity leads to poor runtime and memory performance and eventually to the inability to match, when applied on real-world data. The work presented in this thesis seeks to address these shortcomings. Thereby, we take advantage of the graph structure of metamodels. Consequently, we derive a planar graph edit distance as metamodel similarity metric and mining-based matching to make use of redundant information. We also propose a planar graph-based partitioning to cope with large-scale matching. These techniques are then evaluated using real-world mappings from SAP business integration scenarios and the MDA community. The results demonstrate improvement in quality and managed runtime and memory consumption for large-scale metamodel matching.

Graphene field effect transistors (GFETs) are considered as a candidate for future high-frequency applications. For their realization, the optimal combination of substrate, graphene preparation, and insulator deposition and composition is required.This optimization must be based on an in-depth characterization of the obtained graphene insulator metal (GIM) stack. Hall effect measurements are frequently employed to study such systems, thereby focussing primarily on the charge carrier mobility. In this work we show how an analysis of the sheet Hall coefficient can reveal further important properties of the GIM stack, like, e.g., the interface trap density and the spacial charge inhomogeneity. To that end, we provide an extensive description of the GIM diode, which leads to an accurate calculation of the sheet Hall coefficient dependent on temperature and gate voltage. The gate dependent inverse sheet Hall coefficient is discussed in detail before we introduce the concept of an equivalent temperature, which is a measure of the spacial charge inhomogeneity. In order to test the concept, we apply it to evaluate already measured Hall data taken from the literature. This evaluation allows us to determine the Drude mobility, even at the charge neutrality point, which is inaccessible with a simple one band Hall mobility analysis, and to shed light on the spacial charge inhomogeneity. The formalism is easily adaptable and provides experimentalists a powerful tool for the characterization of their graphene field effect devices.