Search

One of the most important factors influencing the behaviour of a metal when subjected to temperature and pressure is the material's ability to resist deformation. So-called thermomechanical coefficients are widely used for mathematical description of this property. These are the ones that determine the direct effect on deformation resistance. There are temperature coefficient (Kt) deformation degree coefficient (Kɛ) and deformation rate coefficient (Ku) describing the influence of temperature, degree of deformation and deformation rate respectively. The situation becomes complicated because the thermomechanical coefficients are not constants. For example, as the temperature increases, the deformation resistance decreases. Also, the deformation resistance increases with increasing degree of deformation and the deformation rate increases with the deformation resistance increases too. The aim of this paper is to find the most accurate method of interpolating thermomechanical coefficients for various steels and alloys using the least squares method.

Abstract. The paper proposes a binary classification model designed to analyze electroencephalograms data to detecting pathologies associated with epilepsy. The model is based on the Convolutional Neural Network. As input data for the neural network, images obtained by transforming the values of the original electroencephalograms time series based on the Gramian Angular Field matrix were used. The model was trained on data from the Temple University Hospital electroencephalograms Seizure Corpus open data source. The proposed model demonstrated high performance metrics: accuracy – 91%, precision – 92%, recall – 95%, F1-0.93.

The microstrip devices based on multimode resonators represent a class of electromagnetic microwave devices, promising use in tropospheric communication, radar, and navigation systems. The design of wideband bandpass filters, diplexers, and multiplexers with required frequency-selective properties, i.e., bandpass filters, is a complex problem, as electrodynamic modeling is a time-consuming and computationally intensive process. Various planar microstrip resonator topologies can be developed, differing in their topology type, and the search for high-quality structures with unique frequency-selective properties is an important research direction. In this study, we propose an approach for performing an automated search for multimode resonators' conductor topology parameters using a combination of evolutionary computation approach and surrogate modeling. In particular, a variant of differential evolution optimizer is applied, and the model of the target function landscape is built using Gaussian processes. At every iteration of the algorithm, the model is used to search for new high-quality solutions. In addition, a general approach for target function formulation is presented and applied in the proposed approach. The experiments with two microwave filters have demonstrated that the proposed algorithm is capable of solving the problem of tuning two types of topologies, namely three-mode resonators and six-mode resonators, to the required parameters, and the application of surrogated-assisted algorithm has significantly improved overall performance. [ABSTRACT FROM AUTHOR]

The paper presents an approach to the automatic grouping algorithms development based on parametric optimization models for processing high-volume data in the agrarian and industrial complex. Combined search algorithms with alternating randomized neighborhoods show much more stable results (give a smaller minimum value, and also have a low standard deviation of the target function) and hence better performance compared to known (so-called classical) algorithms, such as j-means and k-means.

The article deals with the problem of the formation of a traffic jam on the one of the street Krasnoyarsk cities. According to statistics, the largest set of accidents occur when turning the vehicle. For convenience and traffic priority, traffic lights are set up and additional signals or sections of traffic lights are introduced. In Europe, additional lanes are added, designed separately for turning, where cars can continue to wait for their turn, but if there are a lot of cars, then the additional lane ceases to cope with the flow of cars and begins to load the lane. Another option, when all left turns are prohibited, you need to turn right, where there is one-way traffic, look for a place to turn around and return from another street in order to drive straight to the opposite side of the road where the turn was prohibited. These manipulations can affect the ecology near residential areas and overload small streets, where there is already 1 lane for traffic. The paper proposes the option of transferring the turn to the supermarket to the next intersection in order to increase throughput at large traffic flows, and also uses analytical optimization methods to minimize travel time and maximize passing cars per hour.

In this article, we consider the correction of metric matrices in quasi-Newton methods (QNM) from the perspective of machine learning theory. Based on training information for estimating the matrix of the second derivatives of a function, we formulate a quality functional and minimize it by using gradient machine learning algorithms. We demonstrate that this approach leads us to the well-known ways of updating metric matrices used in QNM. The learning algorithm for finding metric matrices performs minimization along a system of directions, the orthogonality of which determines the convergence rate of the learning process. The degree of learning vectors' orthogonality can be increased both by choosing a QNM and by using additional orthogonalization methods. It has been shown theoretically that the orthogonality degree of learning vectors in the Broyden–Fletcher–Goldfarb–Shanno (BFGS) method is higher than in the Davidon–Fletcher–Powell (DFP) method, which determines the advantage of the BFGS method. In our paper, we discuss some orthogonalization techniques. One of them is to include iterations with orthogonalization or an exact one-dimensional descent. As a result, it is theoretically possible to detect the cumulative effect of reducing the optimization space on quadratic functions. Another way to increase the orthogonality degree of learning vectors at the initial stages of the QNM is a special choice of initial metric matrices. Our computational experiments on problems with a high degree of conditionality have confirmed the stated theoretical assumptions. [ABSTRACT FROM AUTHOR]

For clustering problems based on the model of mixture probability distribution separation, we propose new Variable Neighbourhood Search algorithms (VNS) and evolutionary genetic algorithms (GA) with greedy agglomerative heuristic procedures and compare them with known algorithms. New genetic algorithms implement a global search strategy with the use of a special crossover operator based on greedy agglomerative heuristic procedures in combination with the EM algorithm (Expectation Maximization). In our new VNS algorithms, this combination is used for forming randomized neighbourhoods to search for better solutions. The results of computational experiments made on classical data sets and the testings of production batches of semiconductor devices shipped for the space industry demonstrate that new algorithms allow us to obtain better results, higher values of the log likelihood objective function, in comparison with the EM algorithm and its modifications.

Rapid growth in the volume of transmitted information has lead to the emergence of new wireless networking technologies with variable heterogeneous topologies. With limited radio frequency resources, optimal routing problems arise, both at the network design stage and during its operation. We propose an algorithm based on a minimum loss intensity (greedy-gradient algorithm) to search for optimal routes of information transmission in telecommunication networks. The relevance of the developed algorithm is determined by its practical use in data-transmitting modeling systems. The proposed algorithm satisfies several requirements, such as the speed of the calculations performed, the fulfillment of the conditions for its convergence, and its independence on the selected loss probability function, as well as on the network topology. The idea of the algorithm is a step-by-step recalculation of metrics based on derivatives of the loss intensity function with simultaneous redistribution of information flows along the routes determined by the Floyd algorithm. The comparative efficiency of the proposed algorithm is demonstrated by computational experiments on various network topologies (up to 100 nodes) with various traffic intensities. [ABSTRACT FROM AUTHOR]

Progress in the development of location theory and clustering methods is mainly aimed at improving the performance of algorithms. We study the continuous p-median problem with the Euclidean metric. Various algorithmic combinations of the ALA procedure embedded into more sophisticated algorithmic combinations are used. Two repeated steps of the ALA algorithm solve the simplest discrete optimization problem, and then a series of central point search problems (Weber problem) for each cluster. A parallel implementation of the ALA algorithm is used. The paper studies the issue of choosing from two strategies: solve the Weber problem until accuracy e is achieved at each iteration, or perform only one iteration of the Weiszfeld algorithm at each ALA iteration. [ABSTRACT FROM AUTHOR]