The present work seeks to incorporate a popular descriptive, empirically grounded model of human preference under risk, prospect theory, into the equilibrium theory of noncooperative games. Three primary, candidate definitions are systematically identified on the basis of classical characterizations of Nash Equilibrium; in addition, three equilibrium subtypes are defined for each primary definition, in order to enable modeling of players' reference points as exogenous and fixed, slowly and myopically adaptive, highly flexible and non-myopically adaptive. Each primary equilibrium concept was analyzed both theoretically and empirically; for the theoretical analyses, prospect theory, game theory, and computational complexity theory were all summoned to analysis. In chapter 1, the reader is provided with background on each of these theoretical underpinnings of the current work, the scope of the project is described, and its conclusions briefly summarized. In chapters 2 and 3, each of the three equilibrium concepts is analyzed theoretically, with emphasis placed on issues of classical interest (e.g. existence, dominance, rationalizability) and computational complexity (i.e, assessing how difficult each concept is to apply in algorithmic practice, with particular focus on comparison to classical Nash Equilibrium). This theoretical analysis leads us to discard the first of our three equilibrium concepts as unacceptable. In chapter 4, our remaining two equilibrium concepts are compared empirically, using average-level data originally aggregated from a number of studies by Camerer and Selten and Chmura; the results suggest that PT preferences may improve on the descriptive validity of NE, and pose some interesting questions about the nature of the PT weighting function (2003, Ch. 3). Chapter 5 concludes, systematically summarizes theoretical and empirical differences and similarities between the three equilibrium concepts, and offers some thoughts on future work.

Let there be m parallel machines and n jobs to be scheduled non-preemptively. A job j scheduled on machine i takes p_{i,j} time units to complete, where 1 ≤ i ≤ m and 1 ≤ j ≤ n. For a given schedule, the makespan is the completion time of a machine that finishes last. The goal is to produce a schedule of all n jobs with minimum makespan. This is known as the makespan problem on unrelated parallel machines (UPMs), denoted as R||C_{max}. In this thesis, we focus on subclasses of R||C_{max}. Our research consists of two components. First, a survey of theoretic results for R||C_{max} with a focus on approximation algorithms is presented. Second, we present exact polynomial-time algorithms and approximation algorithms for some subclasses of R||C_{max}. For instance, we present k-approximation algorithms on par with or better than the best known for certain subclasses of R||C_{max}.

This thesis concerns decision procedures for fragments of linear arithmetic and their application to model-checking one-counter automata. The first part of this thesis covers the complexity of decision problems for different types of linear arithmetic, namely the existential subset of the first-order linear theory over the p-adic numbers and the existential subset of Presburger arithmetic with divisibility, with all integer constants and coefficients represented in binary. The most important result of this part is a new upper complexity bound of <b>NEXPTIME</b> for existential Presburger arithmetic with divisibility. The best bound that was known previously was <b>2NEXPTIME</b>, which followed directly from the original proof of decidability of this theory by Lipshitz in 1976. Lipshitz also gave a proof of <b>NP</b>-hardness of the problem in 1981. Our result is the first improvement of the bound since this original description of a decision procedure. Another new result, which is both an important building block in establishing our new upper complexity bound for existential Presburger arithmetic with divisibility and an interesting result in its own right, is that the decision problem for the existential linear theory of the p-adic numbers is in the Counting Hierarchy <b>CH</b>, and thus within <b>PSPACE</b>. The precise complexity of this problem was stated as open by Weispfenning in 1988, who showed that it is in <b>EXPTIME</b> and <b>NP</b>-hard. The second part of this thesis covers two problems concerning one-counter automata. The first problem is an LTL synthesis problem on one-counter automata with integer-valued and parameterised updates and with equality tests. The decidability of this problem was stated as open by G&ouml;ller et al. in 2010. We give a reduction of this problem to the decision problem of a subset of Presburger arithmetic with divisibility with one quantifier alternation and a restriction on existentially quantified variables. A proof of decidability of this theory is currently under review. The final result of this thesis concerns a type of one-counter automata that differs from the previous one in that it allows parameters only on tests, not actions, and it includes both equality and disequality tests on counter values. The decidability of the basic reachability problem on such one-counter automata was stated as an open problem by Demri and Sangnier in 2010. We show that this problem is decidable by a reduction to the decision problem for Presburger arithmetic.

Évaluer la vulnérabilité des réseaux est un enjeu de plus en plus critique. Dans ce mémoire, nous nous penchons sur une approche étudiant la défense d’infrastructures stratégiques contre des attaques malveillantes au travers de problèmes d'optimisations multiniveaux. Plus particulièrement, nous analysons un jeu séquentiel en trois étapes appelé le « Multilevel Critical Node problem » (MCN). Ce jeu voit deux joueurs s'opposer sur un graphe: un attaquant et un défenseur. Le défenseur commence par empêcher préventivement que certains nœuds soient attaqués durant une phase de vaccination. Ensuite, l’attaquant infecte un sous ensemble des nœuds non vaccinés. Finalement, le défenseur réagit avec une stratégie de protection. Dans ce mémoire, nous fournissons les premiers résultats de complexité pour MCN ainsi que ceux de ses sous-jeux. De plus, en considérant les différents cas de graphes unitaires, pondérés ou orientés, nous clarifions la manière dont la complexité de ces problèmes varie. Nos résultats contribuent à élargir les familles de problèmes connus pour être complets pour les classes NP, $\Sigma_2^p$ et $\Sigma_3^p$. Motivés par l’insolubilité intrinsèque de MCN, nous concevons ensuite une heuristique efficace pour le jeu. Nous nous appuyons sur les approches récentes cherchant à apprendre des heuristiques pour des problèmes d’optimisation combinatoire en utilisant l’apprentissage par renforcement et les réseaux de neurones graphiques. Contrairement aux précédents travaux, nous nous intéressons aux situations dans lesquelles de multiples joueurs prennent des décisions de manière séquentielle. En les inscrivant au sein du formalisme d’apprentissage multiagent, nous concevons un algorithme apprenant à résoudre des problèmes d’optimisation combinatoire multiniveaux budgétés opposant deux joueurs dans un jeu à somme nulle sur un graphe. Notre méthode est basée sur un simple curriculum : si un agent sait estimer la valeur d’une instance du problème ayant un budget au plus B, alors résoudre une instance avec budget B+1 peut être fait en temps polynomial quelque soit la direction d’optimisation en regardant la valeur de tous les prochains états possibles. Ainsi, dans une approche ascendante, nous entraînons notre agent sur des jeux de données d’instances résolues heuristiquement avec des budgets de plus en plus grands. Nous rapportons des résultats quasi optimaux sur des graphes de tailles au plus 100 et un temps de résolution divisé par 185 en moyenne comparé au meilleur solutionneur exact pour le MCN. / Evaluating the vulnerability of networks is a problem which has gain momentum in recent decades. In this work, we focus on a Multilevel Programming approach to study the defense of critical infrastructures against malicious attacks. We analyze a three-stage sequential game played in a graph called the Multilevel Critical Node problem (MCN). This game sees two players competing with each other: a defender and an attacker. The defender starts by preventively interdicting nodes from being attacked during what is called a vaccination phase. Then, the attacker infects a subset of non-vaccinated nodes and, finally, the defender reacts with a protection strategy. We provide the first computational complexity results associated with MCN and its subgames. Moreover, by considering unitary, weighted, undirected and directed graphs, we clarify how the theoretical tractability or intractability of those problems vary. Our findings contribute with new NP-complete, $\Sigma_2^p$-complete and $\Sigma_3^p$-complete problems. Motivated by the intrinsic intractability of the MCN, we then design efficient heuristics for the game by building upon the recent approaches seeking to learn heuristics for combinatorial optimization problems through graph neural networks and reinforcement learning. But contrary to previous work, we tackle situations with multiple players taking decisions sequentially. By framing them in a multi-agent reinforcement learning setting, we devise a value-based method to learn to solve multilevel budgeted combinatorial problems involving two players in a zero-sum game over a graph. Our framework is based on a simple curriculum: if an agent knows how to estimate the value of instances with budgets up to B, then solving instances with budget B+1 can be done in polynomial time regardless of the direction of the optimization by checking the value of every possible afterstate. Thus, in a bottom-up approach, we generate datasets of heuristically solved instances with increasingly larger budgets to train our agent. We report results close to optimality on graphs up to 100 nodes and a 185 x speedup on average compared to the quickest exact solver known for the MCN.

This diploma thesis deals with evolutionary algorithms used for travelling salesman problem (TSP). In the first section, there are theoretical foundations of a graph theory and computational complexity theory. Next section contains a description of chosen optimization algorithms. The aim of the diploma thesis is to implement an application that solve TSP using evolutionary algorithms.