|
|
Evolutionary Design of Neural Networks
Kastner, Jan ; Hurta, Martin (referee) ; Sekanina, Lukáš (advisor)
The thesis deals with the implementation of a problem-solving method for the automated design of convolutional neural networks (CNN) architectures. The optimization of two fundamental and often conflicting characteristics, the number of parameters and the quality of CNN classification, is performed using a multi-criteria optimization genetic algorithm (NSGA-II). To encode this problem, the Cartesian genetic programming (CGP) technique is used, which enables the wide range of CNN architectures to be represented, and at the same time, the searched area can be appropriately limited by parameterization. Experiments were performed on the MNIST dataset to understand the effect of population size on the quality of the resulting solution. It is also evident from the results of the experiments that the quality of the architectures found can compete with already established models. This is therefore an alternative approach that does not require human intervention compared to manual design.
|
|
|
Improvement of turbocharger inlet manifold parameters
Kula, Štěpán ; Vacula, Jiří (referee) ; Novotný, Pavel (advisor)
The bachelor thesis deals with the design of a new elbow geometry to reduce the pressure drop in the inlet suction pipe of a turbocharger compressor. At the beginning of the thesis there is a description of the fluid flow and hydraulics losses that result in pressure loss in the pipe. Next, possible methods by which the shape of the piping can be optimized are described. The last, main section deals with the actual optimization of the elbow using the manual method and the Adaptive Single-Objective method, which succeeded in reducing the pressure losses and improving the outlet flow profile.
|
|
|
Optimization of the Manual Coil Wrapping Workplace
Válková, Aneta ; Machačová, Simona (referee) ; Videcká, Zdeňka (advisor)
This bachelor's thesis is focused on the optimization of the manual coil winding workplace in the Elektromotory Drásov spin-off plant of the Innomotics company. There are downtimes at the workplace due to excessive walking, which is the result of an inefficient layout of the workplace. In the theoretical part, individual concepts of process management in production are explained together with methods and tools for improving processes. In the analytical part, the company is described in more detail, especially its current state, including the waste that occurs here. In the design part, two proposals for a new workplace layout are created, which are further compared and, with the help of analyses, their savings in production time, excess walking and financial resources are evaluated.
|
|
|
Advanced Budget Optimization
Golasowski, Martin ; Janíček, Ladislav (referee) ; Popela, Pavel (advisor)
The aim of this diploma thesis is to show possibilities of mathematical modeling of creating budget of public universities in Czech republic based on a deeper study of the issue. In the first part of the work, the financing of universities in the Czech Republic is explained and according to what rules and patterns money is distributed to universities. In the next part, the basic knowledge regarding mathematical optimization and stochastic programming, including the creation of models, is presented. Subsequently, a model of stochastic nonlinear programming is compiled to optimize the distribution of funds from some funding sources to universities. In the last chapter, the model is implemented in the GAMS software and the obtained results are evaluated.
|
|
|
Tax and Accounting Aspects of Emploee Benefits from the Employer's Perspective
Balcárková, Hana ; Gláserová, Jana (referee) ; Křížová, Zuzana (advisor)
The bachelor's thesis is devoted to employee benefits from an accounting and tax point of view in a limited liability company. The first part characterizes the research from the professional literature on the given topic, which the bachelor's thesis deals with. The second part is devoted to the analysis of the current state of provided employee benefits in the company. And as part of the collection of analytical data, a questionnaire survey was conducted in the company. The last part of the work is based on the results of the analytical part and based on these results I suggest possible proposals for improving the current provision of employee benefits, which could increase employee satisfaction in the future.
|
|
|
Initialization for optimization of real world vehicle routing problem
Volf, Robin ; Klhůfek, Jan (referee) ; Sedlák, David (advisor)
The goal of this work is to optimize the initialization phase in solving the rich vehicle routing problem on a specific dataset which comes from a czech logistics company's data. This dataset has characteristics of VRPTW, VRPSPD and HFVRP. A modified cheapest insertion heuristic was chosen for the solution. Cost calculation consists of six partial costs which reflect individual constraints. Proposed heuristic is able to solve the instances in a matter of few seconds. Construction heuristic which generates valid solutions in short time enables more effective optimization in latter parts of solution.
|
|
|
Optimization of Classification Models for Malicious Domain Detection
Pouč, Petr ; Jeřábek, Kamil (referee) ; Hranický, Radek (advisor)
Tato diplomová práce se zaměřuje na rozvoj pokročilých metod pro detekci škodlivých doménových jmen s využitím optimalizačních technik v oblasti strojového učení. Práce zkoumá a hodnotí účinnost různých optimalizačních strategií pro klasifikaci. Jako nástroje pro hodnocení jsem vybral klasifikační algoritmy, které se liší v jejich přístupu, včetně hlubokého učení, techniky rozhodovacích stromů, nebo hledání hyperrovin. Tyto metody byly posouzeny na základě schopnosti efektivně klasifikovat doménová jména v závislosti na použitých optimalizačních technikách. Optimalizace zahrnovala vytvoření přesně označených datových sad, aplikaci technik zpracování dat, pokročilou selekci atributů, řešení nerovnováhy tříd a ladění hyperparametrů. Experimentální část práce prokazuje vynikající úspěšnost kombinováním jednotlivých metod. Přičemž nejlepší modely CNN dosahovaly až 0.9926 F1 při současném snížení FPR na hodnotu 0.300%. Přínos práce spočívá v poskytnutí konkrétních metod a strategií pro efektivní detekci škodlivých doménových jmen v oblasti kybernetické bezpečnosti.
|
|
|
Laser induced breakdown spectroscopy as a technique for elemental analysis of soft biological tissues
Fedra, Vojtěch ; Holub, Daniel (referee) ; Kozáková, Zdenka (advisor)
This bachelor's thesis deals with the optimization of selected measurement parameters for soft tissue analysis using the laser-induced breakdown spectroscopy (LIBS) method. The SNR values (signal-to-noise ratio for spectral ion line Ca II 393,37 nm) were investigated by the LIBS technique for a combination of settings. For the analyses, a laser with a wavelength of 266 nm and a laser with a wavelength of 1064 nm, different values of laser pulse energy, gate delay and for different atmosphere in the ablation chamber (without and with argon purge) were selected. The best results were obtained by using a 1064 nm laser and with argon purge, which led to a significant improvement in SNR. The results of the work are useful for further analysis of soft tissues using the LIBS method.
|
|
|
Information System Assessment and Proposal of ICT Modification
Jančár, Michal ; Koch, Miloš (referee) ; Novák, Lukáš (advisor)
The topic of this bachelor thesis delves into an analysis of business information systems within a selected company and subsequent proposal of changes focused on optimization of business processes. The proposal of changes is based on a theoretical foundation and analyses depicted in this thesis. The proposal of changes was discussed with the management of the company and they will be considered for implementation.
|
|
|
Optimalizace řízení podmínek v inteligentním skleníku
Vilimovský, Dan ; Beran, Vítězslav (referee) ; Bažout, David (advisor)
The work focuses on finding ways to optimize the operation of a smart greenhouse by reducing operating costs and facilitating the sustainability of an ideal environment for plant development in the greenhouse. The aim is to model the conditions in the greenhouse using a developed neural network model that is able to predict the values of environmental parameters such as temperature and humidity. This model is trained on real data obtained from long-term measurements in a test facility. Furthermore, the thesis deals with the possibility of using this neural network model to optimize the control of the greenhouse operation so that ideal development conditions for the plants grown or reduced electricity consumption used for the operation of appliances can be ensured in the future.
|
guest ::
