|
| |
|
| |
|
| |
|
|
Genetic Algorithm Design for Distribution Network Outfits Optimalization
Ondruš, Tomáš ; Skala, Petr (referee) ; Paar, Martin (advisor)
The work deals with genetic algorithms and their potential use in application software to optimize high voltage switching elements of distribution network. Theoretical part explains the basic concepts of genetic algorithms such as a gene, population and chromosome and basic principles of the development of genetic algorithms.. The main task of the thesis is to design the algorithm that will simulate the distribution of the sectionalizers by telecontrolled section switches or reclosers and analyze how to set the the parameters affecting the convergence speed of genetic algorithm. The basic parameters affecting the convergence of breeding, mutation probability, population size or using of elitism. The second goal is finding a suitable set of input parameters for the selected population sizes without and with using elitism. The results of the work determine the most appropriate settings for each generation and determining the approximate number of generations needed to find the best solution. The genetic algorithm applocation was tested on a less extensive distribution network with six switching elements
|
|
|
Evolutionary algorithms
Haupt, Daniel ; Polách, Petr (referee) ; Honzík, Petr (advisor)
The first part of this work deals with the optimization and evolutionary algorithms which are used as a tool to solve complex optimization problems. The discussed algorithms are Differential Evolution, Genetic Algorithm, Simulated Annealing and deterministic non-evolutionary algorithm Taboo Search.. Consequently the discussion is held on the issue of testing the optimization algorithms through the use of the test function gallery and comparison solution all algorithms on Travelling salesman problem. In the second part of this work all above mentioned optimization algorithms are tested on 11 test functions and on three models of placement cities in Travelling salesman problem. Firstly, the experiments are carried out with unlimited number of accesses to the fitness function and secondly with limited number of accesses to the fitness function. All the data are processed statistically and graphically.
|
|
|
Evolutionary computing
Popelka, Jan ; Smékal, Zdeněk (referee) ; Karásek, Jan (advisor)
The aim of this Bachelor's Thesis was to get acquainted with the Evolutionary Optimization Techniques, mainly with the Genetic Algorithm and Genetic Programming. It was subsequently described the role of optimization problem TSP solved using Genetic Algorithms and other Chapter solving Symbolic Regression using Genetic Programming. This optimalization problems were created in the programming JAVA and there are solved practical part of the thesis.
|
|
|
Discovery of Wireless Sensor Network Topology Using Genetic Algorithms
Dalecký, Štěpán ; Samek, Jan (referee) ; Zbořil, František (advisor)
The thesis deals with a design of the genetic algorithm that is able to discover the wireless sensor network topology using signal strength among particular sensors. At first, the thesis describes the theory of genetic algorithm and wireless sensor network. Subsequently, on the basis of this theory, the genetic algorithm serving for the wireless sensor network topology discovery has been designed. The thesis also describes important features of the algorithm implementation. In conclusion, the outcomes have been reviewed.
|
|
|
Methods to detect selection in DNA sequences
Procházka, Ondřej ; Maděránková, Denisa (referee) ; Škutková, Helena (advisor)
The topic of semestral thesis is methods to detect selection in DNA sequences. In the begining of the thesis we will describe molecular evolution. It will be written what made the evolution and how the evolution is shown. Moreover there are gen mutations and mechanisms of diffuse and fixation. It will be defined what pozitive, negative and neutral selection is. The thesis is focused on evolution distance of synonymous and nonsynonymous substitution. There will be described three methods – Nei-Gojobori, Li-Wu-Luo and Comeron. All these methods will be described with mathematic formulas. There will be statistic test to decide what kind of selection ti is – there will be used z-test. In the practical part, there will be information about developed software what counts selection pressure from sequences from databazes in format GenBank and it shows parts where selection is. The software will be used for two data sets with two different genetic codes. The result will be discussed. We will discuss results of all three methods of selection pressure and influence of input parametrs.
|
|
|
Prediction of the Effect of Nucleotide Substitution Using Machine Learning
Šalanda, Ondřej ; Martínek, Tomáš (referee) ; Bendl, Jaroslav (advisor)
This thesis brings a new approach to the prediction of the effect of nucleotide polymorphism on human genome. The main goal is to create a new meta-classifier, which combines predictions of several already implemented software classifiers. The novelty of developed tool lies in using machine learning methods to find consensus over those tools, that would enhance accuracy and versatility of prediction. Final experiments show, that compared to the best integrated tool, the meta-classifier increases the area under ROC curve by 3,4 in average and normalized accuracy is improved by up to 7\,\%. The new classifying service is available at http://ll06.sci.muni.cz:6232/snpeffect/.
|
|
|
Travelling Salesman Problem
Kolář, Adam ; Rozman, Jaroslav (referee) ; Zbořil, František (advisor)
The aim of this bachelor's thesis is to design a testing environment for the traveling salesman problem and compare the effectiveness of different approaches to the solution. The first part discussed the possibility of genetic algorithms, depending on the setting of a crossover, mutations and population size. In the second part, there is the same problem using two types of neural networks. The representative of the self-learning net was chosen Kohonen neural network. Hopfield neural network represents a method of minimizing the energy function with fixed coefficients. At both neural networks, there were described possible advantages and disadvantages. In the end, all the findings were interpreted in a global context.
|
guest ::
