All articles - Robotics

Ikki diapazonli optik modulyatsiya va stimulyatsiya qilingan Brillouin tarqalishiga asoslangan optik vektor analizatori, shuningdek, OptiSystem dasturiy muhitida uni simulyatsiya qilish imkoniyati ko'rib chiqiladi.

The topicality and significance of the subject of dissertation. The rapid growth of the information world rates inextricably leads to increasing levels of complexity while processing information. Insufficiency and lack of the traditional mathematical tools while solving problems of analysis, selecting, classifing and predicting according to data which describe the processes of global social and economic development is becoming obvious. Demand for accuracy and responsiveness of information processing is reinforced in connection with daily increasing information needs of mankind. This encourages scientists to create new approaches to processing large amounts of data with complex structure.
Numerous research works arc implemented to ensure the integration of the republic into the global information space besides implementation of modem information and communication technologies within the demands for the development of social and economic sphere.
Complex integrated systems arc characterized as a large amount of input and output of dates and elements, which have heterogeneous and non-lincar relationships among each other. Information about the system is presented in numerical, qualitative and quantitative forms. As a result, the output of the distribution mechanism of parameters impacting on the system becomes complicated. In some cases, for instance, when severe restrictions on time arc imposed it is impossible to figure out the regularity of relationships.
The rapid development of modem information and communication technologies has led to more effective management of production processes. Due to it material, financial, time and labor costs can be saved. It is one of the reasons which increased scientific and practical interest to data mining in support to decision making. The methods of mining information resources arc the fuzzy inference models, neural networks and hybrid neural, immune, genetic and imitating animal behavior, algorithms of optimization, and generally which embodies the combined model means "Soft Computing" all of which arc based on approximate solutions.
Expanding the scope of the problem in controlling processes, and the development and complexity of the function structure performed by the management requires achieving supporting results in making decision at the level of human opinion. In their turn fuzzy models allow us to describe processes and events of the real world in linguistic terms in natural languages, and the mechanism of fuzzy-decisions is transparent and understandable for people. These obvious advantages extend the capabilities of solving problems in various application fields of science, technology and economics connected with the tasks of analysis, choice, decision-making, classification and prediction in automatic control and monitoring.
The above mentioned statements and highlighted problems justify the actuality of the problem which is the aim of the research meets the challenges of creating systems of data mining models based on the theory of fuzzy sets and reduce errors in alternatives to support decision making by proper selection and configuration of fuzzy model.
This research work is fulfilled to ensure the implementation of that objectives mentioned in the laws of the Republic of Uzbekistan «On Informatization», «On electronic document circulation», the Decree of the President of the Republic of Uzbekistan DP-1989 «On measures for further development of national information and communication system of the Republic of Uzbekistan» dated June 27, 2013 as well as the Decree of the Cabinet of Ministers №355 «On measures of implementation of assess of development state of information and communication technologies in the Republic of Uzbekistan» dated December 31, 2013 to enhance the effectiveness of information in society and the widespread introduction of it in social and economic sectors.
The relevance of the dissertation is characterized by the fact that while managing social and economic processes it is very important to pay a special attention to the development of algorithmic means with powerful mathematical tool that implements the elements of artificial intelligence in solving problems of processing and analysis of dates in the formation of alternatives for decision support with minimum errors.
Purpose of the research is to develop methods, models and algorithmic software of data mining on the basis of mechanism of mathematical tools, for the theory of fuzzy sets for decision-making support of management tasks also the implementation of research results into systems of decision-making for semistructured decisions.
Scientific novelty.
A method of constructing database rules to implement fuzzy model in creating systems mining decision support is established;
A mathematical model of constructing a strategy decision making in semi-logical-linguistic mapping systems to create a system of target monitoring and data mining is developed;
A method of setting the parameters of fuzzy knowledge bases on a modified gradient method and the Markov model to improve the efficiency of the mechanism to support management decision-making system of target monitoring and data mining is formulated;
A description of the methods and data mining models on the basis of fuzzy approach in weakly formalized problems is proposed;
The recurrent equations on the basis of the fuzzy sets theory to find the optimal strategy in fuzzy models in target monitoring arc advanced;
The principles, criteria and requirements for constructing intelligent systems management decision-making based on methods and algorithms of data mining arc established;
Models and algorithms for solving classification and forecasting problems based on data mining in management decisions arc developed
Conclusion
In the dissertation work the system analysis is carried out. Methods, models and algorithms of data mining, realizing an integrated approach to creation of support systems of decision-making arc developed.
The main results of researches arc the following:
1. The analysis of scientific literature, revealing the current state of the problems of building Data Mining systems based on fuzzy-set approach will allow to generate and validate a conceptual framework, design approaches, methods, models and algorithms for constructing decision support systems in targetmonitoring.
2. The mathematical formulation of the problems of decision-making and semi-structured governing of finding optimal solutions to the complex structured tasks thereby provide effictive alternative solutions for the decision support systems.
3. A method of constructing fuzzy inference of model identification in the construction of target monitoring systems and data mining where the model of parametric and structural identification arc implemented. Singleton and Mamdani models have been developed for parametric identification. Model structure identification is implemented on the basis of the algorithms of cluster analysis and subjective methods of separation, the main function of which is to identify the structural characteristics of the fuzzy model in the construction of fuzzy rule base, which serves as a methodological framework for the development of models based on the theory approaches of fuzzy sets.
4. This very method and algorithm of solving the problem of constructing a fuzzy model with the effective implementation of fuzzy rule base justifies the possibility of creating a well-functioning system of data mining.
5. To increase the efficiency of targetmonitoring and data mining while working with a large sets of input data by applying the settings with the fuzzy approach developed gradient method of parametric identification. Implementation of such approaches increases the reliability of decisions of the semistructured tasks by models of fuzzy approach and elements of the fuzzy rule base.
6. Five models of decision- making with the description of fuzzy sets and events for the allowed states of the environment in implementing intelligent systems are proposed. The methods and models display the source data semistructured problems in fuzzy-plural form, evaluating alternatives, searching and finding optimal strategies. A Markov model with fuzzy approach for solving the problem of classification in decision support systems is developed. The given methods and models determine the nature of ill-structured problems, and can improve the accuracy and efficiency of the formation of alternative solutions in systems support decisions.
7. On the basis of the proposed methods and models developed algorithm for problem solving of forecasting the cotton and its application is implemented . The results of prediction of proposed ten step algorithm to construct a fuzzy model arc 0.5-3% more efficient than the existing algorithms. In particular, the accuracy of  the results achieved 96,5-99,8% in problem solving of predicting the yield and cost of raw cotton.
8. An algorithm based on the methods settings fuzzy model for the implementation of decision support is developed and positive results arc justified. Accuracy of objects classification in model problems (IRIS, WINE, «dog-wolf», the diagnosis of cancer) comprises 97-100% in applied problems (forecast yield of raw cotton, cotton varieties selection of conformity, classification workflow system) - 92-98%.
9. The software-oriented to the implementation of decision support systems based on the developed fuzzy set theory approaches have been introduced in the Ministry of Higher and Secondary Special Education of the Republic of Uzbekistan, Department of Agriculture and Water Resources of Dzhizak region, which received acts of implementation and confirms economic benefit from the results of the research done in this dissertation.