Том 78, № 5 (2017)

Consideration was given to a variant of eliminating the mismatch between the dynamic properties of subsystems operating jointly in the complex control systems. An example of realization was given.

We propose ways to synthesize state observers that ensure that the estimation error is bounded on a finite interval with respect to given sets of initial states and admissible trajectories and also simultaneous H_∞-suppression at every time moment of initial deviations and uncertain deviations bounded in L₂-norm, external disturbances for non-autonomous continuous Lipschitz systems. Here the gain of the observers depend on the time and are defined based on a numerical solution of optimization problems with differential linear matrix inequalities or numerical solution of the corresponding matrix comparison system. With the example of a single-link manipulator we show that their application for the state estimating of autonomous systems proves to be more efficient (in terms of convergence time and accuracy of the resulting estimates) as compared to observers with constant coefficients obtained with numerical solutions of optimization problems with linear matrix inequalities.

A method for determination of the time dependence of the anticipated income in the systems of Markov queuing networks with incomes and positive and negative customers was proposed. Subject to this proviso, the incomes from the transitions between the network states are deterministic functions depending on its state and time, and the system incomes in a time unit when they do not change their states depend only on these states. An illustrative example of calculations was given which shows that the anticipated incomes of the network systems can be both increasing and decreasing time functions assuming positive and negative values.

Consideration was given to the control of linear plants under external perturbations and measurement of the quantized plant output. The “consecutive compensator” method was used to design the controller. The obtained algorithm tracks the quantized plant output with respect to the reference signal with precision depending on the quantization step. The simulations illustrate the efficiency of the proposed scheme.

We consider critical sets of an H-regular functional. We propose a condition under which the set of all critical points forms a critical set. We discuss several problems that lead to such sets and show a connection with the notion of Morse index. As examples we consider integral functionals for functions defined on a segment.

We show the results of studying models of the Russian language constructed with recurrent artificial neural networks for systems of automatic recognition of continuous speech. We construct neural network models with different number of elements in the hidden layer and perform linear interpolation of neural network models with the baseline trigram language model. The resulting models were used at the stage of rescoring the N best list. In our experiments on the recognition of continuous Russian speech with extra-large vocabulary (150 thousands of word forms), the relative reduction in the word error rate obtained after rescoring the 50 best list with the neural network language models interpolated with the trigram model was 14%.

Resource-conscious technologies for cutting sheet material include the ICP and ECP technologies that allow for aligning fragments of the contours of cutouts. In this work, we show the mathematical model for the problem of cutting out parts with these technologies and algorithms for finding cutting tool routes that satisfy technological constraints. We give a solution for the problem of representing a cutting plan as a plane graph G = (V,F,E), which is a homeomorphic image of the cutting plan. This has let us formalize technological constraints on the trajectory of cutting the parts according to the cutting plan and propose a series of algorithms for constructing a route in the graph G = (V,F,E), which is an image of an admissible trajectory. Using known coordinates of the preimages of vertices of graph G = (V,F,E) and the locations of fragments of the cutting plan that are preimages of edges of graph G = (V,F,E), the resulting route in the graph G = (V,E) can be interpreted as the cutting tool’s trajectory.

The proposed algorithms for finding routes in a connected graph G have polynomial computational complexity. To find the optimal route in an unconnected graph G, we need to solve, for every dividing face f of graph G, a travelling salesman problem on the set of faces incident to f.

We formulate necessary and sufficient conditions for detecting a fault at the output of a logical element in a combination logical device in a concurrent error-detection system based on Berger’s code. We introduce the notion of a fully testable fault. We show that in order to ensure that all single faults in combination devices are detected with their concurrent checking based on Berger’s code we can use not only the property that it detects 100% of unidirectional errors but also the property that it detects 100% of asymmetrical errors. Due to the latter property, we can reduce structural redundancy of the combination device, when transforming its circuit to one amenable for control, compared to known algorithms for modifying the structure of combination devices into circuits with unidirectionally independent outputs.

This paper systematizes the empirical results on efficiency concepts applied to higher education institutions, data envelopment analysis (DEA) adjusted to heterogeneous samples, inputs and outputs chosen for these institutions and factors tended to make universities efficient. Special attention is paid to the consistency of results yielded by different models.

The problems arising when processing results of foundry sand sieve analysis’ results are formulated. For the solution of the specified problems the algorithm of processing and storage of results of the analysis used in software “Granulometric Calculator” is created.

One-dimensional optimization problems with a polynomial objective function and polynomial matrix inequality constraints are considered. For problems dual to their linear relaxations, a transformation is presented that makes them compatible with the atomic optimization method, both in its basic and in the generalized form with a reduced number of atoms.