|
|
 Выпуск 30
- Blumin S. L. Complete Hypergraphs. Laplasian spectra. Multi-agent systems
Computation of Laplacian spectra is considered for complete and complete uniform undirected and directed hypergraphs which Laplacians are circulants. Standard methods are used for computation of circulant spectra. Applications are discussed to consensus problems in multi-agent systems taking account of agents’ group interactions.
- Saraev P.V. Nonlinear least squares method and block recurrent and iterative procedures in neural networks teaching
The article is devoted to the development of the feedforward neural networks teaching algorithms based on nonlinear least squares method with pseudo-inversion. Application of Cline block pseudo-inversion formula to teaching algorithms based on weights vector decomposition is considered. This approach allows to decrease computational costs by virtue of pseudo-inversion of matrices of small size.
- Sidelnikov J. V., Saltykov S. A. Procedure for selection of the most eligible varieties of the expert methods
In this article the procedure is first developed and justified for selection of the most suitable varieties of expert methods for solving forecasting evaluation tasks, expert method selection recommendations are given for problems of different levels of complexity and extensiveness.
- Yuditsky S. A., Totch D.S. Discrete-continuous model for temporal analysis of network-centric control systems
An integrated model of large systems evolution is suggested. The model consists of two interoperating dynamic network-based submodels: the factors graph with continuous-time factors dynamics, and the actions graph with discrete-time actions dynamics. This combination allows improving the level of adequacy to the real systems compared to the purely discrete model. Time dependency of factors is described in terms of polynomial functions which reveal the interrelation of factors. Action graphs functioning mechanism is improved. A graphical method for the temporal analysis of dynamics in discrete-continuous models is proposed which contributes to the tools for large systems simulation.
- Dorofeyuk Yu. A. Complex control objects identification on the base of piecewise approximation methods
The problem of complex object functioning model development is solved with the aid of algorithms of structure-ranging analysis and piecewise approximation. Two approaches to the solution are considered – the first employs iterative algorithms which implement variational approach to piecewise approximation, while the second uses two-phase algorithms where the processes of input parameter space structuring and local regression model development are separated.
- Andrienko A. Ya., Chadayev A. I. Pilot version of on-board algorithm for on-the-fly recognition of objects
Particular characteristics of the process of stationary and mobile ground targets recognition from the board of pilotless vehicle are considered in reference to onboard recognition algorithms.
- Abdullayeva N. A. Fuzzy approach to povetry level forecasting
Poverty level in Azerbaijan is studied as a function of personal income, unemployment coefficient, inflation rate, living wage and is forecasted to the next three years. So far forecasting problems of this sort were solved with the aim of classical linear regressions. In this article the fuzzy regression model was offered for poverty level forecasting.
- Gorshkova I. V., Klochkov V. V. Economic problems of air transportation network development in under-populated Russian regions
The paper is devoted to specific problems of air transportation network development in regions with low density of population. The influence of low population density on the duration and the cost of air transportation is studied. A complex of organizational, technological decisions and government support decisions aimed at raise of air transportation affordability for under-populated regions’ population is suggested.
- Grishin A. V. Dynamic mathematical modelling of innovative enterprise in conditions of venture financing
We construct a discrete mathematical model of financial and material flows of the company at the stage of production and marketing of innovative products, taking into account various financing options that include leasing and bank credit. The proposed model allows to take into account the processes of functioning of innovative enterprises in sufficient detail, including material, financial, and information flows (the date of raw materials shipment, the date of credit and leasing payments, taxes, etc.), and, compared to similar existing models, allows estimating the impact of a wide range of factors on the financial results of the enterprise under different venture financing options.
- Larionova I. A., Zhaglovskaya A. V.,Boykov A. A., Rozhrov I. M., Markov S. V. Complex management of current assets and their financing sources based on measurement of strategy in use
The problem of measurement and choice of a rational enterprise strategy for current assets and their financing management is considered. The key performance indicator – the integral estimate of the economic status – is built on the basis of financial reports data of the enterprises from the same industry. Strategy assessment method is suggested for current assets management. The method relies on relative amounts of current assets and short-term liabilities. The developed technique is used to show the optimal parameters of the management system that maximize performance.
- Vozhakov A. V., Gitman M. B., Fedoseyev S. A. Complex evaluation for optimal tactical production schedule selection under fuzzy criteria and constraints
The model is considered for evaluation of performance of an automated production scheduling and planning system under fuzzy criteria and constraints. Optimal schedules obtained using different models of integrated estimation are calculated and compared.
- Petrov D. L. Optimal algorithm for data migration in scalable storage clouds
The problem of multi-criteria optimization of data migration schedule in distributed cloud storage is investigated. The algorithm for data migration in scalable cloud storage is developed. We prove that the algorithm has polynomial computational complexity and results in the optimal value of the first criterion.
|
|
