MODIFICATION OF GENETIC ALGORITHM WITH ADAPTIVE CROSSOVER SWITCHING
Y.A. Asanov, S.Y. Beletskaya, Al-Saedi Mohanad Ridha Ganim
The aim of this work is to develop a modification of the adaptive genetic algorithm based on switching crossover in accordance with the degree of elitism of individuals in the population. Despite the enormous amount of research done in the field of evolutionary calculus in recent years, algorithms of this class today have a high prospect of modification. The main aim of research is carried out in order to improve the convergence rate of algorithms (to obtain high-performance optimization methods) and increase the accuracy of the solutions obtained. In the article, for the adaptive tuning of the crossover operator, the concepts of discrete and continuous degree of elitism of individuals are used. In addition, an elitism score is used to adjust the probability of a mutation. This modification has a serious advantage superiority in test problems which are traditionally used to analyze the efficiency of genetic algorithms. The test set used was a quadratic function with three variables, a Rosenbrock function, a step function, a complex fourth-order function with noise, and the Sheckel function. The results of comparing classical genetic algorithms with algorithms using the considered crossover and mutation tuning strategies are presented. An analysis of the results of a computational experiment is presented.
Keywords:genetic algorithm, switching crossover, adaptive mutation tuning, elitism, evolutionary calculus.
APPLICATION OF ARTIFICIAL NEURAL NETWORKS IN THE PROBLEMS OF MANAGING A GENETIC ALGORITHM
D.A. Petrosov, R.A. Vashchenko, A.A. Stepovoi, N.V. Petrosova, A.N. Zelenina
In modern intelligent decision support systems, there is still a problem associated with improving performance in structural and parametric synthesis of large discrete systems with specified behavior based on genetic algorithms. Currently, there are two main areas of research that are designed for mathematical or hardware-based performance improvements. One of the ways to increase hardware performance is to use parallel computing, which includes GPGPU (General-purpose computing on graphics processing units) technology. In this paper, we consider the possibility of increasing the speed of intelligent systems using a mathematical tool of artificial neural networks by introducing a control module for the genetic algorithm directly when performing decision synthesis. The process of structural-parametric synthesis is controlled by predicting and assessing the state of the genetic algorithm (convergence, attenuation, finding the population at local extremes) using artificial neural networks. This allows you to change the parameters of the operators directly in the process of decision synthesis, changing their destructive ability relative to the binary string, which leads to a change in the trajectory of the population in the decision space, and as a result, should increase the speed of intelligent decision support systems.
Keywords: genetic algorithm, intelligent information systems, artificial neural networks, system analysis.
MATHEMATICAL MODEL OF OPTIMIZATION OF THE NETWORK INFRASTRUCTURE OF A DISTRIBUTED ENTERPRISE SYSTEM ON A CLOUD, MISTY AND EDGE TECHNOLOGIES
This article describes the formulation of a mathematical model to optimize the structure of the computer network of a distributed corporate information system based on a multi-level topology. Multi-level topological structure allows you to present the network structure, which includes a detailed description of the relationships between such network objects as active equipment, workstations and servers at different levels of the model, to show as a whole a variety of aspects of data transfer management with support for QoS protocols, technical and software implementation, information space and management. As a result, the design of the computing network of the corporate system can be performed as the development of a single network infrastructure of distributed software and hardware, which includes interconnected and interacting subsystems that solve functional tasks for the management and planning of the enterprise activities within a single approach to achieve the goal of obtaining an optimal architecture and meet the complex system and functional requirements for performance, bandwidth, reliability, etc. The paradigm of cloud, foggy and boundary calculations is used, which allows to build a modern network infrastructure within the framework of the industrial IoT concept. The most typical example of the use of this paradigm can be intelligent power supply networks, distributed monitoring systems for the transportation of gas or oil products, distributed systems of an industrial enterprise. To solve the problem, it is proposed to use an approach based on a modified genetic algorithm. The results of the experiments are presented.
Keywords: multi-level topological structure of distributed corporate system, computer network, cloud computing, fuzzy computing, edge computing, genetic algorithm.
THE SIMULATION OF METAL-DIELECTRIC ANTENNA ON THE BASE OF COMBINED APPROACH
I. Y. Lvovich, A. P. Preobrazhenskiy, O. N. Choporov, E. Ruzhitsky
The structure of modern radio transmitting devices may include antennas formed of both metal and dielectric components-metal-dielectric antennas. They are compact enough and can be placed on various objects of equipment. The paper presents a simulation of a metal-dielectric antenna based on a combined approach. The scheme of antenna construction in different planes is given. The process of scattering of a plane electromagnetic wave on an antenna is considered. The combined algorithm including the method of the integral equation, parallel approach and genetic algorithm is developed. In this paper, the integral equation is used to determine the unknown surface electric currents on the antenna surface, it is solved on the basis of the method of moments. A parallel algorithm was used to speed up the calculations. The impedance matrix is represented as a block matrix. Each block has its own parallel stream. Taking into account the influence of a plane dielectric waveguide on the scattered field, a method associated with a generalized scattering matrix is used. To solve the problem of multi-alternative optimization associated with determining the linear dimensions of the antenna device at a given operating frequency of the antenna, a genetic algorithm is used. As a result, the dimensions of the designed antenna for the specified dimensions of its components are obtained.
Keywords: : antenna, integral equation, parallel approach, optimization, genetic algorithm.
APPLICATION OF FUZZY ARTMAP NETWORK IN INTELLIGENT SYSTEMS OF INVASION DETECTION
The article deals with the organization of intelligent intrusion detection and detection systems. Research in the field of development of information security tools shows that today the most promising and flexible solutions are based on machine learning methods that can prevent damage from intrusions that were not noticed by standard means of combating computer attacks. In the proposed approach, it is proposed to use a sequential reverse search with a return to select significant features and the Fuzzy neural network ARTMAP to detect and diagnose attacks. Network Fuzzy АRTMAP is able to adapt to the dynamics of computer attacks and allows you to recognize intrusions in the information system in real time, without the need to load datasets in batches. This makes it possible to automate the analysis of safety protocols in a continuous mode. The extensive use of ART family networks in intrusion detection tasks makes it possible to consider the search for approaches that improve their performance. In this paper, the control hyperparameters network Fuzzy ARTMAP proposed to adjust automatically with the use of a genetic algorithm According to the results of the computational experiment, the reduced set of characteristics reduces the computation time by 21%. The accuracy of the classification algorithm was 100% and 99.89% for the detection stage and the diagnostic stage, respectively.
Keywords: neural network, Fuzzy ARTMAP, genetic algorithm, intrusion detection, intelligent information security systems.
OPTIMAL NEW CONSUMERS DISTRIBUTION TO URBAN POWER NETWORK SUBSTATIONS
The task of distribution of new power consumers by transformer and distribution substations of the urban power distribution network is considered. The problem is presented in the form of a discrete optimization problem. The article describes three algorithms developed by the authors for solving the problem: a heuristic algorithm of limited search; algorithm that implements methods of genetic search; algorithm based on the construction of Voronoi diagrams. Heuristic algorithm of limited search implements the concept of “greedy” algorithms, where each iteration makes an attempt to connect to the consumer’s network with the least connection costs. In the algorithm that realizes the concept of genetic search, each consumer is assigned one chromosome gene, the allele is the number of the substation to which the connection will be made. In the algorithm based on the construction of Voronoi diagrams, Voronoi diagrams are constructed at each iteration, determining for each substation (transformer or distribution) the set of consumers for which it is the nearest. Comparative analysis of the developed algorithms is carried out with the use of the interactive software complex ELNET. Based on the analysis, a conclusion was made about the efficiency and feasibility of using all the developed algorithms to solve practically significant problems
Keywords: : urban power network, consumers distribution, genetic algorithm, heuristic algorithm, Voronoi diagram.
OPTIMIZATION OF DESIGN TECHNOLOGICAL PROCESSES IN INNOVATIVE ACTIVITY
S.G. Selivanov, O.A. Gavrilova, S.N. Poezjalova, V.V. Nikitin
The article deals with the development of new technological processes, which are created within the framework of innovative projects. Innovative projects include a system of measures that ensure the creation, production and sale of a new type of product or technology for the purpose of generating profit or other useful effects. In the article the concepts “project” and “perspective technological process” are considered, which are close in meaning. A promising technology is introduced without indicating the method or method for implementing the innovation, and the design technology – with its indication by developing the technological part of the new construction, expansion, reconstruction and technical re-equipment projects. Attention is focused on new technological processes, which are developed on the basis of high and critical technologies, and are introduced in production within the framework of innovative projects. The article also discusses the methods of modeling and optimization, which can be used in the development of design technological processes in the machine-building industry. The methods considered are based on the use of artificial intelligence tools, such as genetic and neural network algorithms. The purpose of this publication is a theoretical generalization of methods for optimizing the design of technological processes and the development of new methods for multi-criteria optimization of technological processes based on the methods of artificial intelligence. The practical usefulness of the developments is confirmed by the implementation of projects of technical re-equipment and reconstruction of machine-building production.
Keywords: : electronic document management, electronic document of restricted access, draft of an electronic document of restricted access, creation of a file of a copy of an electronic document of restricted access, protocol of transfer of a file of a copy of an electronic document of restricted access.
RESEARCH OF THE EFFICIENCY OF THE GENETIC ALGORITHM WITH VARIOUS
METHODS OF SELECTION, CROSSOVER TYPES AND STRATEGIES OF FORMATION OF GENERATIONS IN SEARCHING OF EXTREMUMS OF TEST FUNCTIONS
V.S. Maraev, E.A. Bezzubenko, D.A. Cherkashin, A.S. Mihalev
In this article the analysis of materials on genetic algorithms is carried out. The main ideas and principles underlying work of genetic algorithms are considered. The basic stages of the classical genetic algorithm work are analyzed in detail. The review of the most common methods of selection (roulette and tournament), types of crossover (single-point and uniform) and strategies of formation of generations (classical and elitist) is executed.
On test functions the research of genetic algorithm with different methods of selection, types of crossover and strategy of formation of generations is carried out. For each type of algorithm, an estimate of the probability of finding a true solution is given. The received results of the experiments are carefully analyzed. The advantages and disadvantages of different methods of selection, types of crossover, strategy of formation of generations are revealed. The recommendations on the expediency of using genetic algorithms in various situations are stated. The possible directions for further research are defined.
Keywords: genetic algorithm, global extremum, population, generation, selection, crossover, De Jong’s function, Rosenbrock’s function, Rastrigin’s function.
THE PREDICTION OF THE SCATTERING CHARACTERISTICS OF METAL PYRAMIDS IN THE FREQUENCY RANGE BASED ON THE USE OF GENETIC ALGORITHM
A. A. Maksimova, V. N. Kostrova, A. A. Androsov
The analysis model of prediction of radar characteristics is conducting pyramid in the frequency range on the basis of experimental evidence, what is its scattering properties is given. For the case of a vertically incident plane electromagnetic wave using the method of integral equations by numerical way, the calculation of the angular dependences for radar cross section of the pyramid, as well as with the involvement of a genetic algorithm made a prediction of RCS for the specified object in the frequency range is carried out.
Keywords: communication, radio waves, forecasting, optimization, genetic algorithm, integral equations, objects with complex shape.