Le problème de satisfaction de contraintes (CSP) est un problème NP-complet classique en intelligence artificielle qui a suscité un engouement important de la communauté scientifique grâce à la richesse de ses aspects pratiques et théoriques. Cependant, au fil des années un gouffre s'est creusé entre les praticiens, qui développent des méthodes exponentielles mais efficaces pour résoudre des instances industrielles, et les théoriciens qui conçoivent des algorithmes sophistiqués pour résoudre en temps polynomial certaines restrictions de CSP dont l'intérêt pratique n'est pas avéré. Dans cette thèse nous tentons de réconcilier les deux communautés en fournissant des méthodes polynomiales pour tester automatiquement l'appartenance d'une instance de CSP à une sélection de classes traitables majeures. Anticipant la possibilité que les instances réelles ne tombent que rarement dans ces classes traitables, nous analysons également de manière systématique la possibilité de décomposer efficacement une instance en sous-problèmes traitables en utilisant des méthodes de complexité paramétrée. Finalement, nous introduisons un cadre général pour exploiter dans les CSP les idées développées pour la kernelization, un concept fondamental de complexité paramétrée jusqu'ici peu utilisé en pratique. Ce dernier point est appuyé par des expérimentations prometteuses. / The Constraint Satisfaction Problem (CSP) is a fundamental NP-complete problem with many applications in artificial intelligence. This problem has enjoyed considerable scientific attention in the past decades due to its practical usefulness and the deep theoretical questions it relates to. However, there is a wide gap between practitioners, who develop solving techniques that are efficient for industrial instances but exponential in the worst case, and theorists who design sophisticated polynomial-time algorithms for restrictions of CSP defined by certain algebraic properties. In this thesis we attempt to bridge this gap by providing polynomial-time algorithms to test for membership in a selection of major tractable classes. Even if the instance does not belong to one of these classes, we investigate the possibility of decomposing efficiently a CSP instance into tractable subproblems through the lens of parameterized complexity. Finally, we propose a general framework to adapt the concept of kernelization, central to parameterized complexity but hitherto rarely used in practice, to the context of constraint reasoning. Preliminary experiments on this last contribution show promising results.

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / In this work, we present a special class of linear codes: parameterized affine codes. We show that these codes are easy to construct and that given a parameterized affine code one can easily obtain an equivalent projective parameterized code equivalent to it. We also studied some topics which served as the theoretical foundations for the work, such as the theory of Groebner Bases, the footprint of an ideal and some topics of algebraic geometry and commutative algebra. This work has as main goal to obtain the basic parameters (length, dimension and minimum distance) of parameterized codes related and also to relate them to the projective parameterized codes, as done in [7]. We finish by applying the theory of Groebner Bases to the footprint of a certain ideal in order to obtain the basic parameters of the parameterized code over an affine torus. / Neste trabalho apresentamos uma classe especial de códigos lineares: os códigos parametrizados afins. Mostramos que esses códigos são de fácil construção e que, dado um código parametrizado afim, pode-se facilmente obter um código parametrizado projetivo equivalente a ele. Também estudamos algumas teorias que nos serviram como base teórica tais como: a teoria de Bases de Groebner e a Pegada de um ideal e alguns tópicos de geometria algébrica e álgebra comutativa. Este trabalho tem por objetivo principal obter os parâmetros básicos (comprimento, dimensão e distância mínima) dos códigos parametrizados afins e relacioná-los com os códigos parametrizados projetivos, assim como na referência [7]. Encerramos aplicando a teoria de Bases de Groebner a Pegada de um ideal para obter os parâmetros básicos do código parametrizado no toro afim. / Mestre em Matemática

The classical graph edge coloring problem deals in coloring the edges of a given graph with minimum number of colors such that no two adjacent edges in the graph, get the same color in the proposed coloring. In the following work, we look at the other end of the spectrum where in our goal is to maximize the number of colors used for coloring the edges of the graph under some vertex specific constraints. We deal with the MAXIMUM EDGE COLORING problem which is defined as the following –For an integer q ≥2 and a graph G, the goal is to find a coloring of the edges of G with the maximum number of colors such that every vertex of the graph sees at most q colors. The question is very well motivated by the problem of channel assignment in wireless networks. This problem is NP-hard for q ≥ 2, and has been well-studied from the point of view of approximation. This problem has not been studied in the parameterized context before. Hence as a next step, this thesis investigates the parameterized complexity of this problem where the standard parameter is the solution size. The main focus of the work is the special case of q=2 ,i.e. MAXIMUM EDGE 2-COLORING which is theoretically intricate and practically relevant in the wireless networks setting. We first show an exponential kernel for the MAXIMUM EDGE q-COLORING problem where q is a fixed constant and q ≥ 2.We do a more specific analysis for the kernel of the MAXIMUM EDGE 2-COLORING problem. The kernel obtained here is still exponential in size but is better than the kernel obtained for MAXIMUM EDGE q-COLORING problem in case of q=2. We then show a fixed parameter tractable algorithm for the MAXIMUM EDGE 2-COLORING problem with a running time of O*∗(kO(k)).We also show a fixed parameter tractable algorithm for the MAXIMUM EDGE q-COLORING problem with a running time of O∗(kO(qk) qO(k)). The fixed parameter tractability of the dual parametrization of the MAXIMUM EDGE 2-COLORING problem is established by arguing a linear vertex kernel for the problem. We also show that the MAXIMUM EDGE 2-COLORING problem remains hard on graphs where the maximum degree is a constant and also on graphs without cycles of length four. In both these cases, we obtain quadratic kernels. A closely related variant of the problem is the question of MAX EDGE{1,2-}COLORING. For this problem, the vertices in the input graph may have different qε,{1.2} values and the goal is to use at least k colors for the edge coloring of the graph such that every vertex sees at most q colors, where q is either one or two. We show that the MAX EDGE{1,2}-COLORING problem is W[1]-hard on graphs that have no cycles of length four.

Notre travail de recherche s’inscrit dans le cadre du projet de recherche REDiM (Réingénierie des EIAH Dirigée par les Modèles) qui porte sur la réingénierie d'un scénario pédagogique. Il se focalise plus précisément sur l'analyse de traces collectées en session par un EIAH pour fournir à l'enseignant des indicateurs calculés.Dans notre contexte de travail, UTL (Usage Tracking Language) permet de définir des indicateurs sous une forme proche des patrons de conception. Il a été conçu pour répondre aux questions de capitalisation et de réutilisation. Par contre, UTL ne disposait initialement pas de moyens pour spécifier formellement la façon de calculer l’indicateur à partir des traces collectées. De plus, les approches par développement ad hoc d’indicateurs ne permettent pas de modéliser de façon formelle la méthode de calcul. En général, les patrons de conception se limitent à la description, ils ne peuvent donc pas être automatisés. Des descriptions textuelles dans UTL pour produire un indicateur à partir des traces ne permettent pas de générer automatiquement les valeurs d’un indicateur.Notre principal objectif de recherche a donc été de définir des modèles, des méthodes et des outils pour la formalisation et l’automatisation du calcul d’indicateurs. Pour cela, nous avons élaboré une nouvelle version d’UTL qui intègre un langage de combinaison de données nommé DCL4UTL, qui permet de modéliser des indicateurs sous une forme capitalisable, automatisable et réutilisable afin de fournir des indicateurs signifiants à l’enseignant/concepteur. Ces indicateurs peuvent être calculés en temps réel ou après une session, respectivement dans un contexte de tutorat ou de réingénierie du scénario pédagogique.L'originalité de notre approche réside dans le fait que cette version permet non seulement de capitaliser des savoir-faire sur les techniques d'analyse d'usage d'un EIAH, mais aussi, avec le langage DCL4UTL (1) de décrire formellement dans une forme générique des méthodes de modélisation et de calcul d’indicateurs à partir des traces collectées par un EIAH, (2) d’intégrer des fonctions externes (qui proviennent d’autres outils d’analyse), et (3) de créer des données intermédiaires paramétrées facilitant la modélisation et la réutilisation de la méthode de calcul d’indicateurs. Nous avons également développé un outil d’analyse pour calculer les indicateurs modélisés.Cette version est le résultat d'une étude théorique et d’une analyse de l’état de l’art, mais aussi de travaux exploratoires sur la modélisation d’indicateurs et l’analyse de traces. L’approche et le langage ont été validés par plusieurs expérimentations avec plusieurs EIAH existants. / The research topic of this thesis is a part of the REDIM (model driven re-engineering) research project. It focuses specifically on the analysis of tracks collected during the learning session by a TEL (Technology Enhanced Learning) system in order to provide teachers indicators calculated. In our work environment, UTL (Usage Tracking Language) allows users to define the indicators in a form close to the design patterns. It was designed to response capitalization and reuse questions. However, UTL did not initially have any means to formally specify how to calculate indicators based on tracks collected. In general, design patterns are limited to the description, they cannot be automated. In addition, textual descriptions in UTL to produce indicators from tracks do not allow generating automatically an indicator’s values.Our main research objective was therefore to define models, methods and tools for formalizing and automating the calculation of indicators. We propose an extension for UTL named DCL4UTL (Data Combination Language for UTL) to model indicators in a capitalizable, automatable and reusable form to provide meaningful indicators to teachers/designers. With this new version, the indicators can be calculated in real-time or after a learning session in the context of tutoring actions or the reengineering of learning scenarios, respectively.The originality of our approach (DCL4UTL) lies in the fact that this version not only capitalize know-how on analysis techniques of the use an TEL system, but also (1) formally describe models and calculation methods of indicators from tracks collected by a TEL system, (2) integrate external functions (from other analysis tools), and (3) create parameterized intermediate data facilitating the modeling and reuse of indicators’ calculation method. We have also developed an analysis tool to calculate modeled indicators. Our approach and language have been validated by several experiments with several existent TEL systems.

Dans le cadre de cette thèse, nous considérons la complexité paramétrée de problèmes NP-complets. Plus précisément, nous nous intéressons à l'existence d'algorithmes de noyau polynomiaux pour des problèmes d'édition de graphes et de contraintes. Nous introduisons en particulier la notion de branches, qui permet d'obtenir des algorithmes polynomiaux pour des problèmes d'édition de graphes lorsque la classe de graphes cible respecte une décomposition d'adjacence. Cette technique nous permet ainsi d'élaborer les premiers algorithmes de noyaux polynomiaux pour les problèmes Closest 3-Leaf Power, Cograph Edition et Proper Interval Completion. Ces résultats constituent les premiers noyaux polynomiaux pour ces problèmes. Concernant les problèmes d'édition de contraintes, nous étendons la notion de Conflict Packing, qui a déjà été utilisée dans quelques problèmes paramétrés et permet d'élaborer des algorithmes de noyau linéaires pour différents problèmes. Nous présentons un noyau linéaire pour le problème Feedback Arc Set in Tournaments, et adaptons les techniques utilisées pour obtenir un noyau linéaire pour le problème Dense Rooted Triplet Inconsistency. Dans les deux cas, nos résultats améliorent la meilleure borne connue, à savoir un noyau quadratique. Finalement, nous appliquons cette technique sur les problèmes Betweenness in Tournaments et Dense Circular Ordering, obtenant à nouveau des noyaux linéaires, qui constituent les premiers algorithmes de noyau polynomiaux connus pour ces problèmes. / In this thesis, we study the parameterized complexity of several NP-complete problems. More precisely, we study the existence of polynomial kernels for graph and constraints modification problems. In particular, we introduce the concept of branches, which provides polynomial kernels for some graph modification problems when the target graph class admits a so-called adjacency decomposition. This technique allows us to obtain the first known polynomial kernels for the Closest 3-Leaf Power, Cograph Edition and Proper Interval Completion problems. Regarding constraint modification problems, we develop and push further the concept of Conflict Packing, a technique that has already been used in a few parameterized problems and that provides polynomial kernels for several problems. We thus present a linear vertex-kernel for the Feedback Arc Set in Tournaments problem, and adapt these techniques to obtain a linear vertex-kernel for the Dense Rooted Triplet Inconsistency problem as well. In both cases, our results improve the best known bound of $O(k^2)$ vertices. Finally, we apply the Conflict Packing technique on the Betweenness in Tournaments and Dense Circular Ordering problems, obtaining once again linear vertex-kernels. To the best of our knowledge, these results constitute the first known polynomial kernels for these problems.

In this work we develop algorithms for the k-Supplier with Outliers problem. In a network, we are given a set of suppliers and a set of clients. The goal is to choose k suppliers so that the distance between every served client and its nearest supplier is minimized. Clients that are not served are called outliers and the number of allowed outliers is given on input. As k-Supplier with Outliers has numerous applications in logistics, we focus on parameters which are suitable for transportation networks. We study graphs with low highway dimension, which was proposed by Abraham et al. [SODA 2010], and low doubling dimension. It is known that unless P = NP, k-Supplier with Outliers does not admit a (3 − ε)-approximation algorithm for any constant ε > 0. The k-Supplier with Outliers problem is W[1]-hard on graphs of constant doubling dimension for parame- ters k and highway dimension. We overcome both of these barriers through the paradigm of parameterized approximation algorithms. In the case of highway dimension, we develop a (1 + ε)-approximation algorithm for any ε > 0 with running time f(k, p, h, ε) · nO(1) where p is the number of allowed outliers, h is the highway dimension of the input graph, and f is some computable function. In the case of doubling dimension, we develop a (1 + ε)-approximation...

In der klassischen Komplexitätstheorie unterscheiden wir zwischen der Klasse P von in Polynomialzeit lösbaren Problemen, und der Klasse NP-schwer von Problemen bei denen die allgemeine Annahme ist, dass diese nicht in Polynomialzeit lösbar sind. Allerdings sind viele Probleme, die wir lösen möchten, NP-schwer. Gleichzeitig besteht eine große Diskrepanz zwischen den empirisch beobachteten und den festgestellten worst-case Laufzeiten. Es ist bekannt, dass Vorverarbeitung oder Datenreduktion auf realen Instanzen zu Laufzeitverbesserungen führt. Hier stoßen wir an die Grenze der klassischen Komplexitätstheorie. Der Fokus dieser Arbeit liegt auf Vorverarbeitungsalgorithmen für NP-schwere Probleme. Unser Ziel ist es, bestimmte Instanzen eines NP-schweren Problems vorverarbeiten zu können, indem wir die Struktur betrachten. Genauer gesagt, für eine gegebene Instanz und einen zusätzlichen Parameter l, möchten wir in Polynomialzeit eine äquivalente Instanz berechnen, deren Größe und Parameterwert nur durch eine Funktion im Parameterwert l beschränkt ist. In der parametrisierten Komplexitätstheorie heißen diese Algorithmen Kernelisierung. Wir werden drei NP-schwere Graphenprobleme betrachten, nämlich Vertex Cover, Edge Dominating Set und Subset Feedback Vertex Set. Für Vertex Cover werden wir bekannte Ergebnisse für Kernelisierungen vereinheitlichen, wenn der Parameter die Größe einer Entfernungsmenge zu einer gegebenen Graphklasse ist. Anschließend untersuchen wir die Kernelisierbarkeit von Edge Dominating Set. Es stellt sich heraus, dass die Kernelisierbarkeit deutlich komplexer ist. Dennoch klassifizieren wir die Existenz einer polynomiellen Kernelisierung, wenn jeder Graph in der Graphklasse eine disjunkte Vereinigung von konstant großen Komponenten ist. Schließlich betrachten wir das Subset Feedback Vertex Set Problem und zeigen, dass es eine randomisierte polynomielle Kernelisierung hat, wenn der Parameter die Lösungsgröße ist. / In classical complexity theory, we distinguish between the class P, of polynomial-time solvable problems, and the class NP-hard, of problems where the widely-held belief is that we cannot solve these problems in polynomial time. Unfortunately, many of the problems we want to solve are NP-hard. At the same time, there is a large discrepancy between the empirically observed running times and the established worst-case bounds. Using preprocessing or data reductions on real-world instances is known to lead to huge improvements in the running time. Here we come to the limits of classical complexity theory. In this thesis, we focus on preprocessing algorithms for NP-hard problems. Our goal is to find ways to preprocess certain instances of an NP-hard problem by considering the structure of the input instance. More precisely, given an instance and an additional parameter l, we want to compute in polynomial time an equivalent instance whose size and parameter value is bounded by a function in the parameter l only. In the field of parameterized complexity, these algorithms are called kernelizations. We will consider three NP-hard graph problems, namely Vertex Cover, Edge Dominating Set, and Subset Feedback Vertex Set. For Vertex Cover, we will unify known results for kernelizations when parameterized by the size of a deletion set to a specified graph class. Afterwards, we study the existence of polynomial kernelizations for Edge Dominating Set when parameterized by the size of a deletion set to a graph class. We point out that the existence of polynomial kernelizations is much more complicated than for Vertex Cover. Nevertheless, we fully classify the existence of polynomial kernelizations when every graph in the graph class is a disjoint union of constant size components. Finally, we consider graph cut problems, especially the Subset Feedback Vertex Set problem. We show that this problem has a randomized polynomial kernelization when the parameter is the solution size.

Yes / The transfer of suspended sediment can range widely from being diluted to being hyperconcentrated, depending on the local flow and ground conditions. Using the Rouse model and the Kundu and Ghoshal (2017) model, it is possible to look at the sediment distribution for a range of hyper-concentrated and diluted flows. According to the Kundu and Ghoshal model, the sediment flow follows a linear profile for the hyper-concentrated flow regime and a power law applies for the dilute concentrated flow regime. This paper describes these models and how the Kundu and Ghoshal parameters (linear-law coefficients and power-law coefficients) are dependent on sediment flow parameters using machine-learning techniques. The machine-learning models used are XGboost Classifier, Linear Regressor (Ridge), Linear Regressor (Bayesian), K Nearest Neighbours, Decision Tree Regressor, and Support Vector Machines (Regressor). The models were implemented on Google Colab and the models have been applied to determine the relationship between every Kundu and Ghoshal parameter with each sediment flow parameter (mean concentration, Rouse number, and size parameter) for both a linear profile and a power-law profile. The models correctly calculated the suspended sediment profile for a range of flow conditions ( 0.268 𝑚𝑚𝑚𝑚 ≤ 𝑑𝑑50 ≤ 2.29 𝑚𝑚𝑚𝑚, 0.00105 𝑔𝑔 𝑚𝑚𝑚𝑚3 ≤ particle density ≤ 2.65 𝑔𝑔 𝑚𝑚𝑚𝑚3 , 0.197 𝑚𝑚𝑚𝑚 𝑠𝑠 ≤ 𝑣𝑣𝑠𝑠 ≤ 96 𝑚𝑚𝑚𝑚 𝑠𝑠 , 7.16 𝑚𝑚𝑚𝑚 𝑠𝑠 ≤ 𝑢𝑢∗ ≤ 63.3 𝑚𝑚𝑚𝑚 𝑠𝑠 , 0.00042 ≤ 𝑐𝑐̅≤ 0.54), including a range of Rouse numbers (0.0076 ≤ 𝑃𝑃 ≤ 23.5). The models showed particularly good accuracy for testing at low and extremely high concentrations for type I to III profiles.

In der klassischen Komplexitätstheorie werden worst-case Laufzeiten von Algorithmen typischerweise einzig abhängig von der Eingabegröße angegeben. In dem Kontext der parametrisierten Komplexitätstheorie versucht man die Analyse der Laufzeit dahingehend zu verfeinern, dass man zusätzlich zu der Eingabengröße noch einen Parameter berücksichtigt, welcher angibt, wie strukturiert die Eingabe bezüglich einer gewissen Eigenschaft ist. Ein parametrisierter Algorithmus nutzt dann diese beschriebene Struktur aus und erreicht so eine Laufzeit, welche schneller ist als die eines besten unparametrisierten Algorithmus, falls der Parameter klein ist. Der erste Hauptteil dieser Arbeit führt die Forschung in diese Richtung weiter aus und untersucht den Einfluss von verschieden Parametern auf die Laufzeit von bekannten effizient lösbaren Problemen. Einige vorgestellte Algorithmen sind dabei adaptive Algorithmen, was bedeutet, dass die Laufzeit von diesen Algorithmen mit der Laufzeit des besten unparametrisierten Algorithm für den größtmöglichen Parameterwert übereinstimmt und damit theoretisch niemals schlechter als die besten unparametrisierten Algorithmen und übertreffen diese bereits für leicht nichttriviale Parameterwerte. Motiviert durch den allgemeinen Erfolg und der Vielzahl solcher parametrisierten Algorithmen, welche eine vielzahl verschiedener Strukturen ausnutzen, untersuchen wir im zweiten Hauptteil dieser Arbeit, wie man solche unterschiedliche homogene Strukturen zu mehr heterogenen Strukturen vereinen kann. Ausgehend von algebraischen Ausdrücken, welche benutzt werden können, um von Parametern beschriebene Strukturen zu definieren, charakterisieren wir klar und robust heterogene Strukturen und zeigen exemplarisch, wie sich die Parameter tree-depth und modular-width heterogen verbinden lassen. Wir beschreiben dazu effiziente Algorithmen auf heterogenen Strukturen mit Laufzeiten, welche im Spezialfall mit den homogenen Algorithmen übereinstimmen. / In classical complexity theory, the worst-case running times of algorithms depend solely on the size of the input. In parameterized complexity the goal is to refine the analysis of the running time of an algorithm by additionally considering a parameter that measures some kind of structure in the input. A parameterized algorithm then utilizes the structure described by the parameter and achieves a running time that is faster than the best general (unparameterized) algorithm for instances of low parameter value. In the first part of this thesis, we carry forward in this direction and investigate the influence of several parameters on the running times of well-known tractable problems. Several presented algorithms are adaptive algorithms, meaning that they match the running time of a best unparameterized algorithm for worst-case parameter values. Thus, an adaptive parameterized algorithm is asymptotically never worse than the best unparameterized algorithm, while it outperforms the best general algorithm already for slightly non-trivial parameter values. As illustrated in the first part of this thesis, for many problems there exist efficient parameterized algorithms regarding multiple parameters, each describing a different kind of structure. In the second part of this thesis, we explore how to combine such homogeneous structures to more general and heterogeneous structures. Using algebraic expressions, we define new combined graph classes of heterogeneous structure in a clean and robust way, and we showcase this for the heterogeneous merge of the parameters tree-depth and modular-width, by presenting parameterized algorithms on such heterogeneous graph classes and getting running times that match the homogeneous cases throughout.

Die Kernaufgabe der parameterisierten Komplexität ist zu verstehen, wie Eingabestruktur die Problemkomplexität beeinflusst. Wir untersuchen diese Fragestellung aus einer granularen Perspektive und betrachten Problem-Parameter-Kombinationen mit einfach exponentieller Laufzeit, d.h., Laufzeit a^k n^c, wobei n die Eingabegröße ist, k der Parameterwert, und a und c zwei positive Konstanten sind. Unser Ziel ist es, die optimale Laufzeitbasis a für eine gegebene Kombination zu bestimmen. Für viele Zusammenhangsprobleme, wie Connected Vertex Cover oder Connected Dominating Set, ist die optimale Basis bezüglich dem Parameter Baumweite bekannt. Die Baumweite gehört zu der Klasse der Weiteparameter, welche auf natürliche Weise zu Algorithmen mit dem Prinzip der dynamischen Programmierung führen. Im ersten Teil dieser Dissertation untersuchen wir, wie sich die optimale Laufzeitbasis für diverse Zusammenhangsprobleme verändert, wenn wir zu ausdrucksstärkeren Weiteparametern wechseln. Wir entwerfen neue parameterisierte Algorithmen und (bedingte) untere Schranken, um diese optimalen Basen zu bestimmen. Insbesondere zeigen wir für die Parametersequenz Baumweite, modulare Baumweite, und Cliquenweite, dass die optimale Basis von Connected Vertex Cover bei 3 startet, sich erst auf 5 erhöht und dann auf 6, wobei hingegen die optimale Basis von Connected Dominating Set bei 4 startet, erst bei 4 bleibt und sich dann auf 5 erhöht. Im zweiten Teil gehen wir über Weiteparameter hinaus und analysieren restriktivere Arten von Parametern. Für die Baumtiefe entwerfen wir platzsparende Verzweigungsalgorithmen. Die Beweistechniken für untere Schranken bezüglich Weiteparametern übertragen sich nicht zu den restriktiveren Parametern, weshalb nur wenige optimale Laufzeitbasen bekannt sind. Um dies zu beheben untersuchen wir Knotenlöschungsprobleme. Insbesondere zeigen wir, dass die optimale Basis von Odd Cycle Transversal parameterisiert mit einem Modulator zu Baumweite 2 den Wert 3 hat. / The question at the heart of parameterized complexity is how input structure governs the complexity of a problem. We investigate this question from a fine-grained perspective and study problem-parameter-combinations with single-exponential running time, i.e., time a^k n^c, where n is the input size, k the parameter value, and a and c are positive constants. Our goal is to determine the optimal base a for a given combination. For many connectivity problems such as Connected Vertex Cover or Connecting Dominating Set, the optimal base is known relative to treewidth. Treewidth belongs to the class of width parameters, which naturally admit dynamic programming algorithms. In the first part of this thesis, we study how the optimal base changes for these connectivity problems when going to more expressive width parameters. We provide new parameterized dynamic programming algorithms and (conditional) lower bounds to determine the optimal base, in particular, we obtain for the parameter sequence treewidth, modular-treewidth, clique-width that the optimal base for Connected Vertex Cover starts at 3, increases to 5, and then to 6, whereas the optimal base for Connected Dominating Set starts at 4, stays at 4, and then increases to 5. In the second part, we go beyond width parameters and study more restrictive parameterizations like depth parameters and modulators. For treedepth, we design space-efficient branching algorithms. The lower bound techniques for width parameterizations do not carry over to these more restrictive parameterizations and as a result, only a few optimal bases are known. To remedy this, we study standard vertex-deletion problems. In particular, we show that the optimal base of Odd Cycle Transversal parameterized by a modulator to treewidth 2 is 3. Additionally, we show that similar lower bounds can be obtained in the realm of dense graphs by considering modulators consisting of so-called twinclasses.