Bakumenko N. Fuzzy models and evolutionary optimization methods at the airplane engine's creating

The process of the theoretical designing of airplane engines is the object of dissertation research. The aim of research is development of methods and mathematical models necessary for the taking into account qualitative and quantitative descriptions of airplane engine and further using them for the search of its rational parameters. For the solving of the given tasks the methods of fuzzy set theory, that allows to synthesize the fuzzy models of the investigated objects, methods of fuzzy clustering for finding memberships functions parameters, optimization methods, methods of evolutionary search for finding the rational parameters of object, were chosen. The scientific innovation of the got results consists of the following: for the first time the method of acceleration of calculations for the choice of rational parameters of aircraft engines, which provides to carry out the complex calculations on an approximate fuzzy model instead of existent difficult models, that allows to reduce time for rationaldecision choice, saving accuracy. The method of nonlinear dependences modeling by the Mamdani and Takagi-Sugeno fuzzy models at theoretical design of gas-turbine aircraft engines that allows to take into account qualitative descriptions of design object and reduce the design calculation time is improved. The methods of synthesis of the fuzzy Mamdani and Takagi-Sugeno models using fuzzy clustering for finding the memberships functions parameters, that allows to take into account the features of input numeric data and genetic algorithm with the modified mutation operator, that allows to improve the algorithm performance and reduce time for construction of model have been got further development. The practical value of the got results consists of the following: the methods of construction of the fuzzy systems and their programming realization developed in dissertation work allow to take into account expert knowledge in the design process that allows to increase efficiency of optimum parameters search using fuzzy approximation models. The results of dissertation research are used in the Zaporozhian machine-building designer bureau named by A.G. Іvchenko "Progress" and in the educational process of university.