No description available.

In this work, the Traveling Salesman Problem (TSP) is discussed. The Hybrid Genetic Algorithm for solving the TSP is presented. The traveling salesman problem is formulated as follows: given matrix D=(dij)nxn of distances between n objects and the set P of permutations of the integers from 1 to n, find a permutation p=(p(1), p(2), ..., p(n)) P that minimizes. Many heuristic algorithms can be applied for the TSP. Recently, genetic algorithms (GAs) are among the advanced heuristic techniques for the combinatorial problems, like the TSP. genetic algorithms are based on the biological process of natural selection. The original concepts of GAs were developed in 1970s. Many simulations have demonstrated the efficiency of GAs on different optimization problems, among them, bin–packing, generalized assignment problem, graph partitioning, job–shop scheduling problem, set covering problem, vehicle routing. One of the main operators in GAs is the crossover (i.e. solution recombination). This operator plays a very important role by constructing competitive GAs. In this work, we investigate several crossover operators for the TSP, among them, CX (cycle crossover), PMX (partialy mapped crossover), POS (position based crossover), ER (edge recombination crossover), edge-NN (edge recombination crossover, nearest neighbour) and AP (alternating-positions crossover). Comparison of these crossover operators was performed. The results show high efficiency of the edge-NN, ER and PMX crossovers.

Intensity-Duration-Frequency (IDF) curves are among the standard design tools for various engineering applications, such as storm water management systems. The current practice is to use IDF curves based on historical extreme precipitation quantiles. A warming climate, however, might change the extreme precipitation quantiles represented by the IDF curves, emphasizing the need for updating the IDF curves used for the design of urban storm water management systems in different parts of the world, including Canada. This study attempts to construct the future IDF curves for Saskatoon, Canada, under possible climate change scenarios. For this purpose, LARS-WG, a stochastic weather generator, is used to spatially downscale the daily precipitation projected by Global Climate Models (GCMs) from coarse grid resolution to the local point scale. The stochastically downscaled daily precipitation realizations were further disaggregated into ensemble hourly and sub-hourly (as fine as 5-minute) precipitation series, using a disaggregation scheme developed using the K-nearest neighbor (K-NN) technique. This two-stage modeling framework (downscaling to daily, then disaggregating to finer resolutions) is applied to construct the future IDF curves in the city of Saskatoon. The sensitivity of the K-NN disaggregation model to the number of nearest neighbors (i.e. window size) is evaluated during the baseline period (1961-1990). The optimal window size is assigned based on the performance in reproducing the historical IDF curves by the K-NN disaggregation models. Two optimal window sizes are selected for the K-NN hourly and sub-hourly disaggregation models that would be appropriate for the hydrological system of Saskatoon. By using the simulated hourly and sub-hourly precipitation series and the Generalized Extreme Value (GEV) distribution, future changes in the IDF curves and associated uncertainties are quantified using a large ensemble of projections obtained for the Canadian and British GCMs (CanESM2 and HadGEM2-ES) based on three Representative Concentration Pathways; RCP2.6, RCP4.5, and RCP8.5 available from CMIP5 – the most recent product of the Intergovernmental Panel on Climate Change (IPCC). The constructed IDF curves are then compared with the ones constructed using another method based on a genetic programming technique. The results show that the sign and the magnitude of future variations in extreme precipitation quantiles are sensitive to the selection of GCMs and/or RCPs, and the variations seem to become intensified towards the end of the 21st century. Generally, the relative change in precipitation intensities with respect to the historical intensities for CMIP5 climate models (e.g., CanESM2: RCP4.5) is less than those for CMIP3 climate models (e.g., CGCM3.1: B1), which may be due to the inclusion of climate policies (i.e., adaptation and mitigation) in CMIP5 climate models. The two-stage downscaling-disaggregation method enables quantification of uncertainty due to natural internal variability of precipitation, various GCMs and RCPs, and downscaling methods. In general, uncertainty in the projections of future extreme precipitation quantiles increases for short durations and for long return periods. The two-stage method adopted in this study and the GP method reconstruct the historical IDF curves quite successfully during the baseline period (1961-1990); this suggests that these methods can be applied to efficiently construct IDF curves at the local scale under future climate scenarios. The most notable precipitation intensification in Saskatoon is projected to occur with shorter storm duration, up to one hour, and longer return periods of more than 25 years.

This study presents the implementation of Middle East Technical University Neural Network (METU-NN) models for the ionospheric forecasting together with worldwide usage capability of the Internet. Furthermore, an attempt is made to include expert information in the Neural Network (NN) model in the form of neurofuzzy network (NFN). Middle East Technical University Neurofuzzy Network (METU-NFN) modeling approach is developed which is the first attempt of using a neurofuzzy model in the ionospheric forecasting studies. The Web based applications developed in this study have the ability to be customized such that other NN and NFN models including METU-NFN can also be adapted. The NFN models developed in this study are compared with the previously developed and matured METU-NN models. At this very early stage of employing neurofuzzy models in this field, ambitious objectives are not aimed. Applicability of the neurofuzzy systems on the ionospheric forecasting studies is only demonstrated. Training and operating METU-NN and METU-NFN models under equal conditions and with the same data sets, the cross correlation of obtained and measured values are 0.9870 and 0.9086 and the root mean square error (RMSE) values of 1.7425 TECU and 4.7987 TECU are found by operating METU-NN and METU-NFN models respectively. The results obtained by METU-NFN model is close to those found by METU-NN model. These results are reasonable enough to encourage further studies on neurofuzzy models to benefit from expert information. Availability of these models which already attracted intense international attention will greatly help the related scientific circles to use the models. The models can be architecturally constructed, trained and operated on-line. To the best of our knowledge this is the first application that gives the ability of on-line model usage with these features. Applicability of NFN models to the ionospheric forecasting is demonstrated. Having ample flexibility the constructed model enables further developments and improvements. Other neurofuzzy systems in the literature might also lead to better achievements.

The 2019 novel coronavirus has been proven to present several unique features on chest X-rays and CT-scans that distinguish it from imaging of other pulmonary diseases such as bacterial pneumonia and viral pneumonia unrelated to COVID-19. However, the key characteristics of a COVID-19 infection have been proven challenging to detect with the human eye. The aim of this project is to explore if it is possible to distinguish a patient with COVID-19 from a patient who is not suffering from the disease from posteroanterior chest X-ray images using synthetic minority over-sampling and transfer learning. Furthermore, the report will also present the mechanics of COVID-19, the used dataset and models and the validity of the results.

Where there has been a steady development of Optical Character Recognition (OCR) techniques for printed documents, the instruments that provide good quality for hand-written manuscripts by Hand-written Text Recognition  methods (HTR) and transcriptions are still some steps behind. With the main focus on historical ciphers (i.e. encrypted documents from the past with various types of symbol sets), this thesis examines the performance of two machine learning architectures developed within the DECRYPT project framework, a clustering based unsupervised algorithm and a semi-supervised few-shot deep-learning model. Both models are tested on seen and unseen scribes to evaluate the difference in performance and the shortcomings of the two architectures, with the secondary goal of determining the influences of the datasets on the performance. An in-depth analysis of the transcription results is performed with particular focus on the Alchemic and Zodiac symbol sets, with analysis of the model performance relative to character shape and size. The results show the promising performance of Few-Shot architectures when compared to Clustering algorithm, with a respective SER average of 0.336 (0.15 and 0.104 on seen data / 0.754 on unseen data) and 0.596 (0.638 and 0.350 on seen data / 0.8 on unseen data).

The increase of mobile smartphones continues to grow and with it the demand for automation and use of the most offered aspects of the phone, whether in medicine (health care and surveillance) or in user applications (automatic recognition of position, etc.). As part of this work has been created the designs and implementation of the system for the recognition of human activity on the basis of data processing from sensors of smartphones, along with the determination of the optimal parameters, recovery success rate and comparison of individual evaluation. Other benefits include a draft format and displaying numerous training set consisting of real contributions and their manual evaluation. In addition to the main benefits, the software tool was created to allow the validation of the elements of the training set and acquisition of features from this set and software, that is able with the help of deep learning to train models and then test them.

The thesis investigates APT attacks, which are professional targeted attacks that are characterised by long-term duration and use of advanced techniques. The thesis summarises current knowledge about APT attacks and suggests seven symptoms that can be used to check, whether an organization is under an APT attack. Thesis suggests a system for detection of APT attacks based on interaction of those symptoms. This system is elaborated further for detection of attacks in computer networks, where it uses user behaviour modelling for anomaly detection. The detector uses k-nearest neighbors (k-NN) method. The APT attack recognition ability in network environment is verified by implementing and testing this detector.

The thesis is focused on a temperature-rise of a low-voltage switchgear. The first part of the work deals with a technical standard related to the chosen topic, compares software used for designing switchgear in practice and familiarizes with the working environment of the software selected for this work. The main part of the work is dedicated to a measurement and a simulation of the temperature-rise of the low-voltage switchgear. The measurement involved testing of the switchgear with regard to the individual functional units and to the whole switchgear. Subsequently, the simulation of the temperature-rise of the switchgear followed. Software used in this work is SolidWorks Flow Simulation chosen due to a wide range of functions. According to the findings there were proposed adjustments of the switchgear with respect to the technical parameters and the economic aspect in the last part of the thesis.

The project describes the methods useful for observe changes of heart rate in ECG signal. Heart rate variability become (HRV) the conventionally accepted term to describe variations of NN intervals between consecutive heart beats and generally it is function of instantaneous heart rate or NN interval on time. HRV may be evaluated by time domain or frequency domain measures. In Matlab was developed algorithm, realized like function, which counts HRV parameters from ECG signal series. Analysis in time domain adverts to high correlation between statistic and geometric parameters and similarly with signal HRV. Results of frequency domain analysis shows similarity of power spectral density, which was calculated by two different ways (from interpolated and no interpolated signal HRV). Functionality of developed algorithm was verified on each signal. Project results have signification in progress of analysis ECG signal methods with a view to observe pathological changes in heart rate.