abstract: Resource allocation is one of the most challenging issues policy decision makers must address. The objective of this thesis is to explore the resource allocation from an economical perspective, i.e., how to purchase resources in order to satisfy customers' requests. In this thesis, we attend to answer the question: when and how to buy resources to fulfill customers' demands with minimum costs? The first topic studied in this thesis is resource allocation in cloud networks. Cloud computing heralded an era where resources (such as computation and storage) can be scaled up and down elastically and on demand. This flexibility is attractive for its cost effectiveness: the cloud resource price depends on the actual utilization over time. This thesis studies two critical problems in cloud networks, focusing on the economical aspects of the resource allocation in the cloud/virtual networks, and proposes six algorithms to address the resource allocation problems for different discount models. The first problem attends a scenario where the virtual network provider offers different contracts to the service provider. Four algorithms for resource contract migration are proposed under two pricing models: Pay-as-You-Come and Pay-as-You-Go. The second problem explores a scenario where a cloud provider offers k contracts each with a duration and a rate respectively and a customer buys these contracts in order to satisfy its resource demand. This work shows that this problem can be seen as a 2-dimensional generalization of the classic online parking permit problem, and present a k-competitive online algorithm and an optimal online algorithm. The second topic studied in this thesis is to explore how resource allocation and purchasing strategies work in our daily life. For example, is it worth buying a Yoga pass which costs USD 100 for ten entries, although it will expire at the end of this year? Decisions like these are part of our daily life, yet, not much is known today about good online strategies to buy discount vouchers with expiration dates. This work hence introduces a Discount Voucher Purchase Problem (DVPP). It aims to optimize the strategies for buying discount vouchers, i.e., coupons, vouchers, groupons which are valid only during a certain time period. The DVPP comes in three flavors: (1) Once Expire Lose Everything (OELE): Vouchers lose their entire value after expiration. (2) Once Expire Lose Discount (OELD): Vouchers lose their discount value after expiration. (3) Limited Purchasing Window (LPW): Vouchers have the property of OELE and can only be bought during a certain time window. This work explores online algorithms with a provable competitive ratio against a clairvoyant offline algorithm, even in the worst case. In particular, this work makes the following contributions: we present a 4-competitive algorithm for OELE, an 8-competitive algorithm for OELD, and a lower bound for LPW. We also present an optimal offline algorithm for OELE and LPW, and show it is a 2-approximation solution for OELD. / Dissertation/Thesis / Doctoral Dissertation Computer Science 2015

No description available.

Title: Sub-optimal algorithms for solving sliding puzzles Author: Petr Michalík Department: Department of Theoretical Computer Science and Mathematical Logic Supervisor: RNDr. Pavel Surynek, Ph.D. Supervisor's e-mail address: Pavel.Surynek@mff.cuni.cz In the present work techniques for solving the so-called sliding tiles puzzles, which generate optimal or sub-optimal solution, are studied. This thesis focuses especially on a specific variant of the puzzle: the (n^2-1)-puzzle. This work shows and compares current methods for solving this type of problem. A choosen method is a subject to a close analysis of complexity and is also implemented so that theoretical and experimental results could be confronted. An alternative sub-optimal algorithm is proposed and its theoretical analysis is presented. This algorithm is implemented as well and is compared with the existing algorithm. Both the theoretical analysis and the test results show that better (shorter) solutions can often be obtained using this alternative algorithm.

Nous décrivons une procédure numérique pour le recalage d'IRM cérébrales 3D. Le problème d'appariement est abordé à travers la distinction usuelle entre le modèle de déformation et le critère d'appariement. Le modèle de déformation est celui de l'anatomie computationnelle, fondé sur un groupe de difféomorphismes engendrés en intégrant des champs de vecteurs. Le décalage entre les images est évalué en comparant les lignes de niveau de ces images, représentées par un courant différentiel dans le dual d'un espace de champs de vecteurs. Le critère d'appariement obtenu est non local et rapide à calculer. On se place dans l'ensemble des difféomorphismes pour rechercher une déformation reliant les deux images. Pour cela, on minimise le critère en suivant le principe de l'algorithme sous-optimal. L'efficacité de l'algorithme est renforcée par une description eulérienne et périodique du mouvement. L'algorithme est appliqué pour le recalage d'images IRM cérébrale 3d, la procédure numérique menant à ces résultats est intégralement décrite. Nos travaux concernent aussi l'analyse des propriétés de l'algorithme. Pour cela, nous avons simplifié l'équation représentant l'évolution de l'image et étudié l'équation simplifiée en utilisant la théorie des solutions de viscosité. Nous étudions aussi le problème de détection de rupture dans la variance d'un signal aléatoire gaussien. La spécificité de notre modèle vient du cadre infill, ce qui signifie que la distribution des données dépend de la taille de l'échantillon. L'estimateur de l'instant de rupture est défini comme le point maximisant une fonction de contraste. Nous étudions la convergence de cette fonction et ensuite la convergence de l'estimateur associé. L'application la plus directe concerne l'estimation de changement dans le paramètre de Hurst d'un mouvement brownien fractionnaire. L'estimateur dépend d'un paramètre p > 0 et nos résultats montrent qu'il peut être intéressant de choisir p < 2. / We present a numerical procedure for the matching of 3D MRI. The problem of image matching is addressed through the usual distinction between the deformation model and the matching criterion. The deformation model is based on the theory of computational anatomy and the set of deformations is a group of diffeomorphisms generated by integrating vector fields. The discrepancy between the two images is evaluated through comparisons of level lines represented by a differential current in the dual of a space of vector fields. This representation leads to a quickly computable non-local criterion. Then, the optimisation method is based on the minimization of the criterion following the idea of the so-called sub-optimal algorithm. We take advantage of the eulerian and periodical description of the algorithm to get an efficient numerical procedure. This algorithm can be used to deal with 3d MR images and numerical experiences are presented. In an other part, we focus on theoretical properties of the algorithm. We begin by simplifying the equation representing the evolution of the deformed image and we use the theory of viscosity solutions to study the simplified equation. The second issue we are interested in is the change-point estimation for a gaussian sequence with change in the variance parameter. The main feature of our model is that we work with infill data and the nature of the data can evolve jointly with the size of the sample. The usual approach suggests to introduce a contrast function and using the point of its maximum as a change-point estimator. We first get an information about the asymptotic fluctuations of the contrast function around its mean function. Then, we focus on the change-point estimator and more precisely on the convergence of this estimator. The most direct application concerns the detection of change in the Hurst parameter of a fractional brownian motion. The estimator depends on a parameter p > 0, generalizing the usual choice p = 2. We present some results illustrating the advantage of a parameter p < 2.