1 documents found
Information × Registration Number 0210U001274, 0107U000598 , R & D reports Title Development Theoretical Bases of Neural Networks of an Adaptive Resonance for Optimization Complex Processes popup.stage_title Head Dmitrienko Valery, Registration Date 02-02-2010 Organization National technical university "Kharkiv polytechnic institute" popup.description2 A research object is processes of creation on the basis of onemodul and hierarchical neural networks (NS) of adaptive resonance theory (ART) the systems of support making decision (SPPR) for optimization difficult dynamic objects and processes in the conditions of a priori substantial vagueness. A purpose of work is perfection on the basis of development of theory of onemodul and hierarchical NS ART methods of creation SPPR for optimization of difficult objects and processes in the conditions of a priori substantial vagueness (on the examples of processes of edge processing optical polymers (LOOP) and management mobile composition with a hauling asynchronous drive (TAPAS)). Research methods are a theory of continuous and discrete neuron networks of adaptive resonance, theory of LOOP, theory of TAPAS. First on the basis of development of theory of onemodul and hierarchical NS ART with compact storage of information, with the great numbers of possible decisions, with new architectures and algorithms of teaching for the decision task of determination optimum structure of technological process making wares of polymeric optics SPPR is developed for optimization processes of LOOP. Results correspond a world level, as raising got further development task of optimum management the operations of edge processing which unlike known takes into account operating limitations in quality of polymeric wares and current wear of instrument. First in Ukraine developed systems of support making decision a machinist (SPPRM), using the electronic maps of motion to en-route and repeatedly conducted hauling calculations, allowing in real time and depending on a change a travelling and weather situation to correct strategy conduct mobile composition on the set route. The method of hauling calculations is improved, that allowed to promote exactness of calculations of basic parameters motion of composition. Developed SPPRM does not yield to the best foreign developments. Thus for development base of knowledges SPPRM annex storage was first used on NS ART with three entrances, built on the basis of continuous NS ART with the new method of compact storage of information. Development of such NS ART corresponds a world level, as new networks open possibility of creation of NS and bases of knowledges with the great numbers of associations. The expected results are opened possibilities of decision wide class of tasks: designs which arise up in industry at a management different technological processes; tasks of electromagnetic compatibility of difficult objects (in particular, aircrafts) and defence them from the powerful electromagnetic impulses of natural and artificial origin; management and control of mobile composition of the Ukrainian railways; automations of medical diagnostics; developments of the systems of support of making decision at edge processing of optical polymeric wares. Product Description popup.authors Баленко Олексій Іванович Гладких Тетяна Валентинівна Даниленко Олександр Федорович Заковоротний Олександр Юрійович Заполовський Микола Йосипович Ключкевич Євгеній Коц Володимир Ліпчанський Максим Валентинович Леонов Сергій Юрійович Мезенцев Микола Вікторович Носков Валентин Іванович Петрушанський Максим Пилипенко Оксана Валентинівна Поворознюк Анатолій Іванович Поворознюк Оксана Анатоієвна Попенко Артем Серков Олександр Анатолійович Хавіна Інна Петрівна popup.nrat_date 2020-04-02 Close
R & D report
Head: Dmitrienko Valery. Development Theoretical Bases of Neural Networks of an Adaptive Resonance for Optimization Complex Processes. (popup.stage: ). National technical university "Kharkiv polytechnic institute". № 0210U001274
1 documents found

Updated: 2026-03-22