The approach of system improvement, methodology and information technology of turbofan engine elements main parameters and operating tolerances robust estimation on basis of inverse problems concept by stochastic optimization problem are proposed. The method of multicriterion system modification problem quasisolution searching with input data uncertainly and with bordering of feasible solution class is offered. Quasisolution synthesis is realized by regularization of smoothing functional minimum searching with using the A.N. Tihonov’s method. The multicriterion modification problem numerical solution searching evolution method was created, it based on genetic algorithm using. On basis of development methodology the system aerodynamic turbofan engine (TFE) compressor improvement were made.

When projecting the gas turbine an important problem is an ensuring the high values of gas turbine parameters and required gas turbine operating characteristics on the different operating conditions. These requirements can be reached by engine function units system perfecting on base of multicriterion stochastic optimization problems solution. Three stochastic optimization problems definitions were formulated. Each problem has own features and can be used for different application solution. These applied problems are: M-problem can be used on the technical system unit conceptual design stage; V-problem can be used for the problem solution of tolerancing during the technical system unit production; P-problem can be used for interval analysis of technical system functional unit. The multicriterion stochastic optimization problem rational decision is realized by the evolutional method. This method makes it possible to find the solution with given accuracy by attraction the less information recourses than standard methods. In the stochastic optimization problems definitions the input data random character is taken into account. It makes it possible to find the optimal values of desired parameters. These parameters ensure the maximal probability of finding the objective function in given range.