L.Е. Mistrov – Dr.Sc. (Eng.), Associate Professor, Professor,

Air Force Research Center «VVA», Central Branch of the Russian State Unitary Enterprise (Voronezh) E-mail: mistrov_le@mail.ru 

Е.М. Shepovalov – Flight Simulator Head,

Air Force Research Center «VVA» (Voronezh)

E-mail: shepovaloff2011@yandex.ru

The current stage of training of aircraft crews is associated with widespread use at the stage of ground training of aviation multifunction simulators (AMT). The basis of their application is the solution of a variety of diverse educational problems based on the additive combination of various types of elementary technological operations (IT). IT, representing a variety of geometric shapes with different spatio-temporal parameters, when forming training tasks based on the analysis of their content, can be repeated, rotated and / or branched. This causes significant difficulties in the formation of IT options for the implementation of AMT training tasks, necessitating the development of analysis methods and the selection of preferred options.

Assuming the existence of monotonicity in each coordinate and convexity in their totality of IT options, we consider a linear and nonlinear form of the utility function. For the linear form of the utility function, based on the preferences of the decision makers (groups of experts) and their competence, a maximum likelihood function is determined that allows ordering IT options by decision makers (DM). If the values of the estimates of alternative IT options are significantly different, then the assumption of the linearity of the function of their utility is rejected. In the non-linear form of the utility function, on the set of IT options, the options with the maximum and minimum values of the utility function are found, the distance between them is determined, and the options with the minimum distance between the compared options are removed from consideration on the basis of dividing the set of IT options into the subsets of the best and worst, while the power of the resulting subsets will not be sufficient for a direct analysis of the decisionmaker. After that, on the subsets of the best and worst cases of IT, a discriminating hyperplane is built, parallel to these subsets, cutting off the share of the worst options based on the normal of the obtained hyperplanes.

If it is impossible to compare the DM decision among themselves of the IT options, their utility functions become fuzzy defined, and the comparison results are also fuzzy. Their determination is carried out on the basis of the introduction of the membership function, which is an extension of the averaging function for deterministic extrapolation. The parity of this ambiguity is carried out by additionally taking into account the expert competence function, which allows us to represent the membership estimate in the form of a piecewise-linear function, the range of values of which is found from solving a system of nonlinear equalities with a small number of compared IT options. With a large number of IT options, the search for a solution is carried out on the basis of modeling using the Monte Carlo method, which allows ordering options on a quantitative scale of quality criteria for this decision-maker taking into  account the degree of their confidence and choosing the optimal options for implementing AMT training tasks on a set of preferred options.

Mistrov L.E., Shepovalov E.M. Method for forming options of elementary technological operations for the implementation of educational problems by the aviation multi-functional simulator. Achievements of modern radioelectronics. 2020. V. 74. № 9. P. 18–25. DOI: 10.18127/j20700784-202009-02. [in Russian]

1. Mistrov L.E., Shepovalov E.M. Osnovy prinyatiya resheniy po razrabotke (modernizatsii) aviatsionnykh mnogofunktsional'nykh trenazherov. Uspekhi sovremennoy radioelektroniki. 2019. № 6. S. 66–76. DOI: 10.18127/j20700784-201906-08. [in Russian]
2. Devyatov D.B., Sysoev V.V., Chirko M.S. Sravnitel'nyy analiz metodov ekstrapolyatsii ekspertnykh otsenok na osnove printsipa maksimal'nogo pravdopodobiya. Nadezhnost' i kontrol' kachestva. 1984. № 12. S. 12–15. [in Russian]
3. Bard Y. Nelineynoe otsenivanie parametrov. M.: Statistika. 1979. [in Russian]
4. Afifi A., Eyden S. Statisticheskiy analiz. Podkhod s ispol'zovaniem EVM. M.: Mir. 1982. [in Russian]
5. Nechetkie mnozhestva v modulyakh upravleniya i iskusstvennogo intellekta. Pod red. D.A. Pospelova. M.: Nauka. Gl. red. fiz.-mat. literatury. 1986. [in Russian]