RNA secondary structure is very important in many biological processes. Efficient structure prediction can give information for experimental investigations of these processes. Many available programs for secondary structure prediction exist. Some of them use single sequence, the others use more related sequences. Pseudoknots are still problematic for most methods. This work presents several methods and publicly available software and the implementation of minimum free energy method is described.

In this paper we discuss the inverse problem of the identification of mechanical stresses by measuring the deformation of an electric potential field in a so called differential strain gauge (D-DMS). We derive a mathematical model, where the forward operator is given in terms of an elliptic boundary value problem. Derivatives of the forward operator are considered and the solution of the inverse problem via a least-squares minimization is introduced. Here, the discretized problem is solved with the Gauss-Newton method. Numerical studies of practical interest are presented.

Liver perfusion can be modelled by Darcy's flow in multiple connected com- partments. The first part of the present thesis shows in detail the existence of a solution to the multi-compartmental model. The flow in each compartment in this model is characterized by a permeability tensor, which is obtained from the geometry of liver vasculature. It turns out that this tensor might be singular, which potentially causes solvability problems. The second part deals with this abnormality in one compartment. By using the theory of degenerate Sobolev spaces, an appropriate weak formulation is defined. Analogues of Poincar'e and traces inequalities in this degenerate setting are proved, which also imply the existence of the weak solutions. In addition, this part justifies another possibil- ity how to deal with degenerate permeability, which is regularizing the tensor by adding a small isotropic permeability to it. In the third part, the aim is to find subdomains of autonomous perfusion with respect to the source positions. This is formulated as a minimization problem and several numerical results are presented. 1

General game playing is an area of artificial intelligence which focuses on creating agents capable of playing many games from some class. The agents receive the rules just before the match and therefore cannot be specialized for each game. Deepstack is the first artificial intelligence to beat professional human players in heads-up no-limit Texas hold'em poker. While it is specialized for poker, at its core is a general algorithm for playing two-player zero-sum games with imperfect information - continual resolving. In this thesis we introduce a general version of continual resolving and compare its performance against Online Outcome Sampling Monte Carlo Counterfactual Regret Minimization in several games.

The thesis addresses time-domain inverse electromagneticscattering for determining unknown characteristics of an objectfrom observations of the scattered .eld. Applications includenon-destructive characterization of media and optimization ofmaterial properties, for example the design of radar absorbingmaterials.A nother interesting application is the parameteroptimization of subcell models to avoid detailed modeling ofcomplex geometries. The inverse problem is formulated as an optimal controlproblem where the cost function to be minimized is thedi.erence between the estimated and observed .elds, and thecontrol parameters are the unknown object characteristics. Theproblem is solved in a deterministic gradient-basedoptimization algorithm using a parallel 2D FDTD scheme for thedirect problem.This approach is computationally intensive sincethe direct problem needs to be solved in every optimizationiteration in order to compute an estimated .eld.H ighlyaccurate analytical gradients are computed from the adjointformulation.In addition to giving better accuracy than .nitedi.erences, the analytical gradients also have the advantage ofonly requiring one direct and one adjoint problem to be solvedregardless of the number of parameters. When absorbing boundary conditions are used to truncate thecomputational domain, the equations are non-reversible and theentire time-history of the direct solution needs to be storedfor the gradient computation.Ho wever, using an additionaldirect simulation and a restart procedure it is possible tokeep the storage at an acceptable level. The inverse method has been successfully applied to a widerange of industrial problems within the European project,IMPACT (Inverse Methods for Wave Propagation Applications inTime-Domain).T he results presented here includecharacterization of layered dispersive media, determination ofparameters in subcell models for thin sheets and narrow slotsand optimization problems where the observed .eld is given bydesign objectives.

With the increasing capacity of today's integrated circuits, a number ofheterogeneous  system-on-chip (SoC)  architectures  in embedded  systemshave been proposed. In order to achieve energy and design cost efficientstreaming applications  on these  systems, new design  space explorationframeworks  and  performance  analysis  approaches are  required.   Thisthesis  considers three state-of-the-art  SoCs architectures,  i.e., themulti-processor SoCs (MPSoCs)  with network-on-chip (NoC) communication,the hybrid CPU/FPGA architectures, and the run-time reconfigurable (RTR)FPGAs.  The main topic of the  author?s research is to model and capturethe  application  scheduling,  architecture  customization,  and  bufferdimensioning  problems, according to  the real-time  requirement.  Sincethese  problems  are NP-complete,  heuristic  algorithms and  constraintprogramming solver are used to compute a solution.For  NoC  communication  based  MPSoCs,  an  approach  to  optimize  thereal-time    streaming    applications    with   customized    processorvoltage-frequency levels and memory  sizes is presented. A multi-clockedsynchronous  model  of  computation   (MoC)  framework  is  proposed  inheterogeneous  timing analysis and  energy estimation.   Using heuristicsearching  (i.e., greedy  and  taboo search),  the  experiments show  anenergy reduction (up to 21%)  without any loss in application throughputcompared with an ad-hoc approach.On hybrid CPU/FPGA architectures,  the buffer minimization scheduling ofreal-time streaming  applications is addressed.  Based  on event models,the  problem  has  been  formalized  decoratively  as  constraint  basescheduling,  and  solved  by  public domain  constraint  solver  Gecode.Compared  with  traditional  PAPS  method,  the  proposed  method  needssignificantly smaller  buffers (2.4%  of PAPS in  the best  case), whilehigh throughput guarantees can still be achieved.Furthermore, a  novel compile-time analysis approach  based on iterativetiming  phases is  proposed  for run-time  reconfigurations in  adaptivereal-time   streaming   applications  on   RTR   FPGAs.   Finally,   thereconfigurations analysis and design trade-offs analysis capabilities ofthe proposed  framework have been  exemplified with experiments  on bothexample and industrial applications. / Andres

This dissertation is a complete unifying approach to the fundamentals, industrial applications and software implementation of an important branch of process-engineering principles and practice, called process integration. The latter refers to the system-oriented, thermodynamically-based and integrated approaches to the analysis, synthesis and retrofit of process plants, focusing on integrating the use of materials and energy, and minimizing the generation of emissions and wastes. This work extends process integration to include applications for industrial water reuse and wastewater minimization and presents previous developments in a unified manner. The basic ideas of process integration are: (1) to consider first the big picture by looking at the entire manufacturing process as an integrated system; (2) to apply process-engineering principles to key process steps to establish a priori targets for the use of materials and energy, and for the generation of emissions and wastes; and (3) to finalize the details of the process design and retrofit later to support the integrated view, particularly in meeting the established targets. Pinch technology is a set of primarily graphical tools for analyzing a process plant's potential for energy conservation, emission reduction and waste minimization. Here, we identify targets for the minimum consumption of heating and cooling utilities, mass-separating agents, freshwater consumption, wastewater generation and effluent treatment and propose economical grassroots designs and retrofit projects to meet these goals. An emerging alternative approach to pinch technology, especially when analyzing complex water-using operations and effluent-treatment systems, is mathematical optimization. We solve nonlinear programming problems for simple water-using operations through readily available commercial software. However, more complex, nonconvex problems require sophisticated reformulation techniques to guarantee optimality and are the subject of continuing academic and commercial development. This work develops the principles and practice of an environmentally significant breakthrough of process integration, called water-pinch technology. The new technology enables the practicing engineers to maximize water reuse, reduce wastewater generation, and minimize effluent treatment through pinch technology and mathematical optimization. It applies the technology in an industrial water-reuse demonstration project in a petrochemical complex in Taiwan, increasing the average water reuse (and thus reducing the wastewater treatment) in the five manufacturing facilities from 18.6% to 37%. This dissertation presents complete conceptual and software developments to unify the known branches of process integration, such as heat and mass integration, and wastewater minimization, and explores new frontiers of applications to greatly simplify the tools of process integration for practicing engineers. / Ph. D.

Dans cette thèse, nous étudions des stratégies d’allocation séquentielle de ressources. Le modèle statistique adopté dans ce cadre est celui du bandit stochastique à plusieurs bras. Dans ce modèle, lorsqu’un agent tire un bras du bandit, il reçoit pour récompense une réalisation d’une distribution de probabilité associée au bras. Nous nous intéressons à deux problèmes de bandit différents : la maximisation de la somme des récompenses et l’identification des meilleurs bras (où l’agent cherche à identifier le ou les bras conduisant à la meilleure récompense moyenne, sans subir de perte lorsqu’il tire un «mauvais» bras). Nous nous attachons à proposer pour ces deux objectifs des stratégies de tirage des bras, aussi appelées algorithmes de bandit, que l’on peut qualifier d’optimales. La maximisation des récompenses est équivalente à la minimisation d’une quantité appelée regret. Grâce à une borne inférieure asymptotique sur le regret d’une stratégie uniformément efficace établie par Lai et Robbins, on peut définir la notion d’algorithme asymptotiquement optimal comme un algorithme dont le regret atteint cette borne inférieure. Dans cette thèse, nous proposons pour deux algorithmes d’inspiration bayésienne, Bayes-UCB et Thompson Sampling, une analyse à temps fini dans le cadre des modèles de bandit à récompenses binaires, c’est-à-dire une majoration non asymptotique de leur regret. Cette majoration permetd’établir l’optimalité asymptotique des deux algorithmes. Dans le cadre de l’identification des meilleurs bras, on peut chercher à déterminer le nombre total d’échantillons des bras nécessaires pour identifier, avec forte probabilité, le ou les meilleurs bras, sans la contrainte de maximiser la somme des observations. Nous définissons deux termes de complexité pour l’identification des meilleurs bras dans deux cadres considérés dans la littérature, qui correspondent à un budget fixé ou à un niveau de confiance fixé. Nous proposons de nouvelles bornes inférieures sur ces complexités, et nous analysons de nouveaux algorithmes, dont certains atteignent les bornes inférieures dans des cas particuliers de modèles de bandit à deux bras, et peuvent donc être qualifiés d’optimaux. / In this thesis, we study strategies for sequential resource allocation, under the so-called stochastic multi-armed bandit model. In this model, when an agent draws an arm, he receives as a reward a realization from a probability distribution associated to the arm. In this document, we consider two different bandit problems. In the reward maximization objective, the agent aims at maximizing the sum of rewards obtained during his interaction with the bandit, whereas in the best arm identification objective, his goal is to find the set of m best arms (i.e. arms with highest mean reward), without suffering a loss when drawing ‘bad’ arms. For these two objectives, we propose strategies, also called bandit algorithms, that are optimal (or close to optimal), in a sense precised below. Maximizing the sum of rewards is equivalent to minimizing a quantity called regret. Thanks to an asymptotic lower bound on the regret of any uniformly efficient algorithm given by Lai and Robbins, one can define asymptotically optimal algorithms as algorithms whose regret reaches this lower bound. In this thesis, we propose, for two Bayesian algorithms, Bayes-UCB and Thompson Sampling, a finite-time analysis, that is a non-asymptotic upper bound on their regret, in the particular case of bandits with binary rewards. This upper bound allows to establish the asymptotic optimality of both algorithms. In the best arm identification framework, a possible goal is to determine the number of samples of the armsneeded to identify, with high probability, the set of m best arms. We define a notion of complexity for best arm identification in two different settings considered in the literature: the fixed-budget and fixed-confidence settings. We provide new lower bounds on these complexity terms and we analyse new algorithms, some of which reach the lower bound in particular cases of two-armed bandit models and are therefore optimal

Background: Cannabis drugs are the most used illicit drug in the Czech Republic. About 1,78 million people use cannabis, most of them are young people aged 15-34 years (Mravčík et al., 2020). Use in the young age, regular and intensive use of large amounts of cannabis is associated with the negative impact on health and life of users. Previous studies have described the harm redcution strategies, less is known about relative occurrence of the hram reduction stategies, especially in the czech environment. Aims: The aim of the study was to investigate what kind of harm reduction strategies are used by cannabis users. Methods: The research was conducted as a quantitative study. This mapping study was aimed to describe the behaviour of cannabis users in relation to use of the harm reduction strategies, and to explore the relative occurrence of these strategies. As a method of data collection was used a questionnaire survey. Results: Most frequently emerged harm reduction strategies among cannabis users are in relation to the effects of use on physical health. Most of these strategies focus on minimising the harms associated with smoking cannabis, particularly marijuana cigarettes. Mental health strategies are based on the concept of set, where users often do not use if they observe negative feelings...

The diploma thesis deals with the simulation of induction motor control with respect to high efficiency. The theory of an induction motor is described here, with emphasis on its losses. Scalar and vector control are also described here. Vector control is optimized for higher efficiency. Subsequently, the creation of a model in the program MATLAB - Simulink is described here, for the comparison of vector control with and without optimization.