Search

В обучении искусственной нейронной сети одной из центральных проблем является начальная инициализация и настройка весовых коэффициентов, связанная с псевдослучайной инициализацией весовых коэффициентов. В статье описан базовый генетический алгоритм, а также метод определения весовых коэффициентов с использованием данного алгоритма. Также представлен комбинированный метод определения весовых коэффициентов, предусматривающий на первом этапе начальную инициализацию с помощью генетического алгоритма и использование стохастического градиентного спуска на втором этапе обучения Предлагаемые методы проверены на ряде искусственных нейронных сетей прямого распространения для различных задач по бинарной классификации реальных и синтетических данных, а также для однозначной многоклассовой классификации рукописных цифр на изображениях из базы данных MNIST. Искусственные нейронные сети сконструированы на основании теоремы Колмогорова-Арнольда. В данной статье проведен сравнительный анализ двух методов определения весовых коэффициентов – с использованием генетического алгоритма и градиентного спуска. По результатам сравнительного анализа сделан вывод о возможности применения генетического алгоритма для определения весовых коэффициентов как в качестве алгоритма начальной инициализации искусственной нейронной сети, так и в качестве алгоритма настройки весовых коэффициентов

Introduction/purpose: Tracked vehicles play a vital role across various domains, from military operations to construction and agriculture. This study focuses on improving the efficiency of tracked vehicles by optimizing both gear ratios and gear-shifting strategies while preserving other performance aspects. Methods: The optimization process involves a genetic algorithm for determining optimal gear ratios, considering performance constraints. Furthermore, the paper introduces a gear-shifting optimization algorithm aimed at enhancing fuel economy to the maximum, while allowing for a valid comparison between two sets of gear ratios. Results: Optimizing gear ratios leads to substantial reductions in fuel consumption, as the engine operates within more efficient regions. Additionally, the optimized gear-shifting strategy further enhances efficiency, resulting in a fuel consumption reduction exceeding 12%, when combined with the optimized gear ratios. Conclusions: This paper offers a direct and robust approach for optimizing powertrain gear ratios and gear-shifting strategies in tracked vehicles. The results demonstrate significant improvements in fuel efficiency without compromising other critical vehicle performance parameters. [ABSTRACT FROM AUTHOR]

Рассматривается использование различных типов эвристических алгоритмов на основе технологий мягких вычислений для распределения задач в группах мобильных роботов, выполняющих односложные операции в едином рабочем пространстве: генетические алгоритмы, муравьиные алгоритмы и искусственные нейронные сети. Показано, что данная задача является NP-сложной и ее решение прямым перебором для большого числа заданий невозможно. Исходная задача сведена к типовым NP-полным задачам: обобщенной задаче поиска оптимальной группы замкнутых маршрутов от одного депо и задаче коммивояжера. Представлены описание каждого из выбранных алгоритмов и сравнение их характеристик. Приводится пошаговый алгоритм работы с учетом выбранных генетических операторов и их параметров при заданном объеме популяции. Представлена общая структура разработанного алгоритма, позволяющего достаточно эффективно решить многокритериальную оптимизационную задачу с учетом временных затрат и интегрального критерия эффективности роботов, учитывающего энергетические затраты, функциональную насыщенность каждого агента группы и т.д. Показана возможность решения исходной задачи с использованием муравьиного алгоритма и обобщенного поиска оптимальной группы замкнутых маршрутов. Для многокритериальной оптимизации показана возможность линейной свертки полученного векторного критерия оптимальности за счет введения дополнительных параметров, характеризующих групповое управление: общее КПД функционирования всех роботов, затраты энергии на функционирование группы поддержки и энергия на размещение одного робота на рабочем поле. Для решения задачи распределения заданий с использованием нейронной сети Хопфилда произведено ее представление в виде графа, полученного в ходе перехода от обобщенной задачи поиска оптимальной группы замкнутых маршрутов от одного депо к задаче коммивояжера. Показателем качества выбран суммарный путь, пройденный каждым из роботов группы.

Накопление данных о процессах управления проектами и типовых решениях сделало актуальными исследования, связанные с применением методов инженерии знаний для многокритериального поиска вариантов, которые задают оптимальные настройки параметров проектной среды. Цель: разработка методики поиска и визуализации групп проектов, которые могут быть оценены на основе концепции доминирования и интерпретироваться в терминах проектных переменных и показателей эффективности. Методы: обогащение выборки с сохранением неявной связи между проектными переменными и показателями эффективности осуществляется с помощью прогнозирующей нейросетевой модели. Для обнаружения фронта Парето в многомерном критериальном пространстве используется набор генетических алгоритмов. Онтология проектов определяется после кластеризации вариантов в пространстве решений и преобразования кластерной структуры в критериальное пространство. Автоматизация поиска в многомерном пространстве зоны наибольшей кривизны фронта Парето, определяющей равновесные проектные решения, их визуализация и интерпретация осуществляются с помощью плоского дерева решений. Результаты: плоское дерево строится при любой размерности критериального пространства и имеет структуру, которая имеет топологическое соответствие с проекциями разделяемых образов кластеров из многомерного пространства на плоскость. Для различных видов преобразований и корреляций между показателями эффективности и проектными переменными показано, что участки наибольшей кривизны фронта Парето определяются либо содержимым целого кластера, либо частью вариантов, представляющих “лучший” кластер. Если на плоском дереве к правому верхнему углу примыкает неразделенный прямоугольник кластера, то его представители в критериальном пространстве хорошо отделены от остальных кластеров и при максимизации показателей эффективности наиболее приближены к идеальной точке. Все представители такого кластера являются эффективными решениями. Если кластер-победитель содержит внутри дерева решений доминируемые варианты, то “лучший” кластер представляют оставшиеся варианты, которые задают оптимальные настройки проектных переменных. Практическая значимость: предложенная методика поиска и визуализации групп проектов может найти применение при выборе условий ресурсного и организационно-экономического моделирования проектной среды, обеспечивающих оптимизацию рисков, стоимостных, функциональных и временных критериев.

В работе рассматривается метод расчёта кинематики шарнирной конструкции и применение генетического алгоритма для обеспечения требуемой траектории движения заданных точек конструкции. Сначала описаны типы звеньев произвольных шарнирных конструкций, их математическое описание и построение на их основе систем нелинейных уравнений, описывающих кинематику шарнирных конструкций. Полученные нелинейные системы решаются многократным применением метода Ньютона для систем нелинейных уравнений. Построенный алгоритм применяется для расчёта кинематики шарнирной конструкции, представляющей собой модель конечности «кинетических скульптур» Тео Янсена. Целью расчёта является подбор оптимальных параметров конечности, которые позволят применить её для модели робота с четырьмя опорными конечностями. Строится функция приспособленности траектории опорной точки и методом генетического отбора подбирается лучшая траектория, которая обеспечивает прямолинейное движение корпуса робота относительно земли. Теоретические расчёты подкрепляются численным моделированием и визуализацией. В конце статьи приводится график, иллюстрирующий успешное применение разработанной модели. На основе теоретических построений создан реальный прототип шагающего робота с четырьмя конечностями методом 3D печати комплектующих и последующей сборки.

Copyright of Cybernetics & Systems Analysis / Kibernetiki i Sistemnyj Analiz is the property of V.M. Glushkov Institute of Cybernetics of NAS of Ukraine and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Radio Electronics, Computer Science, Control is the property of Zaporizhzhia National Technical University and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

In given work is given benchmark analysis drawn near algorithm of the decision for popular problem of the theories of the timetables at condition of homogeneity instrument information handling.

This paper gives the estimation of the tournament selection, used in the genetic algorithm, for the uniform instruments of information processing, via the comparison of the standard diagram of tournament selection with its modifications, with the aid of the computational experiment.

Copyright of System Research & Information Technologies / Sistemnì Doslìdžennâ ta Ìnformacìjnì Tehnologìï is the property of Institute for the Applied System Analysis at the NTUU KPI and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Electrical Engineering & Electromechanics is the property of National Technical University, Kharkiv Polytechnic Institute and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Scientific Bulletin of National Mining University is the property of National Mining University, State Higher Educational Institution and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

In the known methods of symbolical regression by search of the solution with the help of a genetic algorithm, there is a problem of crossover. Genetic programming performs a crossover only in certainpoints. Grammatical evolution often corrects a code after a crossover. Other methods of symbolical regression use excess elements in a code for elimination of this shortcoming. The work presents a new method of symbolic regression on base of binary computing trees. The method has no problems with a crossover. Method use a coding in the form of a set of integer numbers like analytic programming. The work describes the new method and some examples of codding for mathematical expressions.

Introduction. A quite general class of practical tasks is guided by the set covering problem: schedules building, layout of service stations, and creation of electronic circuits. It defines relevance of searching methods to improve the solution efficiency of this task. Materials and Methods. Techniques of the set covering problem solution by exact and approximate algorithms are considered. The genetic algorithm is used as the approximate method, and the branch and bounds algorithm - as the exact method. Research Results. The genetic algorithm in all its modifications on time response characteristics has shown predictability and stability in all series of experiments. The branch and bounds method was applied to the set covering task, and it has shown exact results. Discussion and Conclusions . The conducted research shows that for small sets, it is expedient to use the branch and bounds method which has demonstrated fast runtime with an assured exact result. For large sets, it is recommended to use the genetic algorithm which guarantees receiving a result with a negligible error where the execution time shift is stable and predictable.

Stochastic optimization methods such as genetic algorithm, particle swarm optimization algorithm, and others are successfully used to solve optimization problems. They are all based on similar ideas and need minimal adaptation when being implemented. But several factors complicate the application of stochastic search methods in practice: multimodality of the objective function, optimization with constraints, finding the best parameter configuration of the algorithm, the increasing of the searching space, etc. This paper proposes a new Cascade Object Optimization and Modification Algorithm (COOMA) which develops the best ideas of known stochastic optimization methods and can be applied to a wide variety of real-world problems described in the terms of object-oriented models with practically any types of parameters, variables, and associations between objects. The objects of different classes are organized in pools and pools form the hierarchical structure according to the associations between classes. The algorithm is also executed according to the pool structure: the methods of the upper-level pools before changing their objects call the analogous methods of all their subpools. The algorithm starts with initialization step and then passes through a number of iterations during which the objects are modified until the stop criteria are satisfied. The objects are modified using movement, replication and mutation operations. Two-level version of COOMA realizes a built-in self-adaptive mechanism. The optimization statistics for a number of test problems shows that COOMA is able to solve multi-level problems (with objects of different associated classes), problems with multimodal fitness functions and systems of constraints. COOMA source code on Java is available on request.

Copyright of Radio Electronics, Computer Science, Control is the property of Zaporizhzhia National Technical University and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Radio Electronics, Computer Science, Control is the property of Zaporizhzhia National Technical University and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

This paper considers the problem of optimization of the characteristics of scattering of electromagnetic waves on periodic electrodynamic structure. The solution of the scattering problem is based on the method of integral equations, the optimization of the characteristics is based on the genetic algorithm. Recommendations on the parameters of the periodic structure under given angles are given.

In this article has been identified linear stationary roll dynamic model for unmanned air vehicle with the weak exciting input signal and sensor measurements noise using the two-step method, maximum likelihood method and the genetic optimization algorithm. Due to the weak frequency content of the input signal, eigenvalues of the information matrix are close to zero, and the use of the output error method often gives the wrong solutions in each cycle of the identification algorithm based on the Monte Carlo method. Ill-conditioned information matrix in the identification of dynamic model occurs due to linear relationships between variables. The two-step identification method finds the ratio of the parameters in the first stage. In the second step, the identification algorithm is conducted for the vector of parameters with reduced size. The two-stage identification method gives the right solution with the permissible errors for each execution of the identification process

This paper discusses the simulation of scattering of radio waves diffraction on the antenna. The scattering characteristics are determined on the basis of the method of integral equations. The optimization was performed for characteristics based on genetic algorithm.

The weighed and unweighted minimal set-cover problem, its applicability for the solution of the major optimization practical tasks, such as arrangement of service points, assignment of crews in transport, as well as the integrated-circuit and conveyer lines designing is considered. The paper objective is to describe methods of improving efficiency of this task solution. The principle of operation of the genetic algorithm and the applicability of its modification as a method of the solution to the set-cover problem are defined. Greedy strategy of Chvatal for the set-cover problem solution is considered. An exhaustive algorithm as an exact algorithm for the solution of small-size tasks is developed. The modified genetic algorithm developed by Nguyen M. H. is described. A software tool for comparing the performance of these algorithms is created. It is concluded that the solution of the set-cover problem by our genetic algorithm modification is more effective than by the genetic algorithm of Nguyen M. H. or the greedy strategy. And the results obtained in small-size tasks are noted for insignificant error.

Эволюционные вычисления в интерактивной музыке В интерактивной музыке компьютер, как правило, не просто компонент единой системы, но активное действующее лицо, зачастую - равноправный исполнитель. В то же время его функционирование предопределено вычислительным алгоритмом и конкретной программой, которая в него заложена исследователем/исполнителем/программистом. Стремление придать машине большую творческую независимость ведет к вовлечению обширного исследовательского поля искусственного интеллекта в пространство музыкального искусства. Эволюционные вычисления, генетический алгоритм н генетическое програм мирование, являясь частью этого поля, позволяют находить решения в тех областях музыкальной практики, где точный ответ невозможен в силу специфики задачи (например, при неопределенности конечного результата как такового). Гибкость, предоставляемая эволюционными вычислениями, является одним из привлекательных моментов, однако рождает и ряд проблем, связанных как с выбором подходящего для решения музыкально-исследовательской задачи набора геномов, так и с определением функции приспособленности. Проблема, в особенности, возникает, когда мы выходим из теоретической, исследовательской области в сферу музыкального творчества. Очевидно, если пространство эволюционных вычислений не будет управляться набором тех или иных музыкальных правил, стохастическая во многом природа генетических процессов может привести к полной неуправляемости конечного результата: иногда революционно нового, но, чаще, бессмысленного и художественно малозначимого.

The work objective is to study a critical traveling salesman problem which is NP complicated task of the discrete optimization. It is shown that only heuristics is appropriate in achieving this goal. The result of the ant colony algorithm (ACA) and genetic algorithm (GA) sharing is presented for the problem solution. The point is that the problem is solved using only mutations of various types (without crossover). The investigated GA is improved by the elitist strategy. The testing is done on graphs of the middle and large dimension. An “elite” sample obtained by the ACA is improved by a mean of 11%. It is shown that the efficiency of the genetic algorithm depends directly on the number of ants in the generation, and on the number of algorithm iterations. Target function values are improved more than twofold after the introduction of the elitist strategy. Increasing the number of ACA runs raises the efficiency of the algorithm by approximately 2%.

Modeling of the non-destructive testing system of defects in solids is performed. Specifically, the inverse geometric problems of the elasticity theory for a flat rectangular area on reconstructing circular cavities and cracks breaking the body surface are considered. Additional information for solving these problems is a setting of the first four natural resonance frequencies. The inverse problem solution is based on the minimization of the residual functional between the measured input source information and the data calculated during the numerical solution of direct problems with the given parameters of defects. As a tool for solving direct problems, the finite element method implemented in FlexPDE program is used. The functional minimization is carried out by using a genetic algorithm (GA) implemented in the developed GAFEMNDT program. The program algorithm and GA settings used in the numerical experiments are described. The experiments results on determining parameters of defects (coordinates of centre, radius, coordinates of surface cracking and its size) are presented. The results demonstrate adequacy of the additional information to overcome the problem ill-posedness, as well as high efficiency of the proposed algorithm both in accuracy of detecting defects parameters, and in their search speed.

The paper focuses on the problem of control system synthesis and a numerical method of the network operator is proposed to search a solution. Apart from a well-known approach when the stated control synthesis problem is solved directly by the method of network operator considering given criteria and terminal conditions, the present paper describes a two-stage synthesis. The first stage is supposed to solve numerically a set of optimal control tasks for different initial values from some given space. The second stage makes an approximation of the previously received optimal trajectories by means of the network operator. An example illustrates the two-stage synthesis of a control system for a spaceship descent to the Moon

Feasibility of genetic search algorithms application for implementation of the cryptanalysis of block cipher methods is considered. A distinctive feature of the bioinspired cryptanalysis methods application (in particular, genetic methods) is the possibility of using the encryption (or decryption) algorithm as an objective function for the suitability evaluation of the key defined by genetic operations. Consequently, when using the bioinspired cryptanalysis methods, the S key definition (for example, when using type 2 cryptanalysis) depends not so much on the complexity of the ciphering transformations, as on the bioinspired method which is to provide a sufficient variety of key generation that shows the significance of the research task of the bioinspired algorithms feasibility (in particular, genetic search methods) for the block cryptosystem cryptanalysis. It is noted also that as the distinctive feature of both block cipher methods, and the genetic algorithms is their internal parallelism, then the task of developing a cryptanalysis algorithm based on the parallel implementation of the constituent stages is relevant. An algorithm of the block methods cryptanalysis on the example of the DES standard on the basis of its parallel version is offered; the experiment results of the quasioptimal key determination obtained at the parallel algorithm implementation on the 8-letter text blocks are given. It is noted that time costs of the algorithm realization do not exceed the time of the known cryptanalysis implementation.

Copyright of Electronics & Control Systems is the property of National Aviation University and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Electronics & Control Systems is the property of National Aviation University and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Cybernetics & Systems Analysis / Kibernetiki i Sistemnyj Analiz is the property of V.M. Glushkov Institute of Cybernetics of NAS of Ukraine and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Problems of Economy is the property of Research Centre for Industrial Developmen Problems of Nas and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

The computing circuit for search the optimal size/shape of structural design to improve the bearing capacity of metal trusses based on the evolutionary modeling is developed. We considered strengthening by increasing the rod cross section and a change structural layout of the object. The topology formation is performed based on exception of some redundant rods of structure and a change position of nodes. By increasing of cross section rods the strengthening is modeled by introducing finite elements with hard consoles. Optimization is carried out by path the minimum cost of structures. The examples of strengthen in truss equipment building overhead cranes and increase operational load on the roof is given.

Copyright of Neuro-Fuzzy Modeling Techniques in Economics is the property of Kyiv National Economic University named after Vadym Hetman and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Neuro-Fuzzy Modeling Techniques in Economics is the property of Kyiv National Economic University named after Vadym Hetman and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Machinery & Energetics is the property of National University of Life & Environmental Sciences of Ukraine and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Eastern-European Journal of Enterprise Technologies is the property of PC TECHNOLOGY CENTER and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Electrical Engineering & Electromechanics is the property of National Technical University, Kharkiv Polytechnic Institute and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Journal of Nano- & Electronic Physics is the property of Sumy State University and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

The article justifies a necessity of early identification of crisis situations at by-product coke plants. It develops a scheme of interconnection of main components of identification of crisis situations, which consists of three main blocks: formation of the system of identification indicators, classification of financial and economic situations of a company and automatic recognition of the situation class. The following methods of the fuzzy logic were used as instruments of realisation of the above listed blocks: the method of fuzzy clusterisation of c-averages for classification of financial and economic situations and fuzzy neural networks for recognition. The article also offers an integral indicator built upon the matrix of degrees of belonging of situations to clusters, obtained with the help of the method of fuzzy c-averages. The indicator was used in the models of recognition of crisis situations. Moreover, the article offers a two-stage algorithm of recognition of crisis situations, which includes a preliminary synthesis of fuzzy rules on the basis of the method of c-averages with further education of the obtained fuzzy model of the Sugeno type by the genetic algorithm.В статье обоснована необходимость ранней идентификации кризисных ситуаций на предприятиях коксохимической промышленности. Разработана схема взаимосвязи основных составляющих идентификации кризисных ситуаций, которая состоит из трех основных блоков: формирование системы показателей идентификации, классификация финансово-хозяйственных ситуаций предприятия и автоматическое распознавание класса ситуаций. В качестве инструментария реал��зации вышеперечисленных блоков использовались методы нечёткой логики, а именно: для классификации финансово-хозяйственных ситуаций – метод нечеткой кластеризации c-средних, для распознавания – нечеткие нейронные сети. Также в статье был предложен интегральный показатель, построенный на основе матрицы степеней принадлежности ситуаций кластерам, полученной с помощью метода нечетких с-средних. Показатель был использован в моделях распознавания кризисных ситуаций. Кроме того, был предложен двухэтапный алгоритм распознавания кризисных ситуаций, который включает в себя предварительный синтез нечетких правил на основе метода c-средних с дальнейшим обучением полученной нечеткой модели типа Сугено генетическим алгоритмом.

An iterative optimization scheme of metal beam structures is developed with using a genetic algorithm with alternative ways of forming populations and adaptive mutation operator, crossover and inversion. This algorithm allows solving large problems to avoid occurrence of "empty" populations, has a relatively low complexity and provides opportunity to obtain new rational decisions. The

In the article, the methods of vibrations and dynamic stability analysis of bars having the nonlinear diagram deformation of material are suggested. The example of calculation of a bar with hinged supports of ends is given. The graphs of bar deflection as a function of time parameter are constructed.

The problem of synthesis of control system is considered. It is shown that classical statement of the problem does not provide satisfactory quality of control. The corrected statement of the problem where solution guaranties high quality of control for any initial conditions is proposed.

The method of fuzzy control system parameters synthesis and adjustment is given. The method uses genetic programming elements. Fuzzy regulator is included in direct control circuit of nonlinear dynamic object.