PROBLEMS OF INFORMATION TRANSMISSION
A translation of Problemy Peredachi Informatsii


Volume 5, Number 3, July–September, 1969
Back to contents page

CONTENTS                   Powered by MathJax

 

On the Logical Foundations of Information Theory and Probability Theory
A. N. Kolmogorov
pp. 1–4

Abstract—New logical foundations are proposed for information theory and probability theory; they are based on the concept of the complexity of a sequence.

 

The Condition for Informational Regularity in Stationary Gaussian Processes
I. A. Ibragimov
pp. 5–16

Abstract—A stationary process $\xi(t)$ is informationally regular if the quantity of information in the past $\{\xi(t),\:t\le 0\}$, relative to the future $\{\xi(t),\:t\ge\tau\}$, tends to zero as $\tau\to\infty$. A description is given of the spectral characteristics of informationally regular Gaussian stationary processes.

 

Conditions for Boundedness of Information on a Gaussian Process Contained in Another Process
D. S. Apokorin
pp. 17–21

Abstract—The necessary and sufficient conditions are given for which the mutual information between two Gaussian processes is finite. More effective sufficiency conditions are calculated for certain special cases.

 

Constant-Weight Codes and Tactical Configurations
N. V. Semakov and V. A. Zinoviev
pp. 22–28

Abstract—We consider the general properties of constant-weight codes with the structure of tactical configurations and prove that the upper bounds for the number of words, which follow from Jonson’s estimates, are achieved only for such codes. We give the distribution of distances in complete constant-weight codes which are isomorphic with Steiner systems and quasi-complete codes which are close to them. We also construct a new family of quasi-complete constant-weight codes.

 

Complexity of the Combinational Unit under Noise-Immune Coding of the States of an Automaton
Yu. L. Sagalovich
pp. 29–34

Abstract—Two methods of organizing the combinational unit of an automaton are discussed under noise-immune coding of the (complete or only the internal) states of an automaton. For each method estimates are given for the complexity of the combinational unit, and the limits of applicability are given.

 

Local Criteria for Stable Choice of Power Level
V. L. Stefanyuk
pp. 35–46

Abstract—When radiated-power levels are selected for a group of mutually interfering radio stations, it is no longer possible to use the natural criterion for a single channel, the a priori specified signal-to-noise ratio, since independent power control at each transmitter would lead to unstable operation in certain cases. Certain classes of criteria are considered that always provide stable selection of  “reasonable” singal-to-noise ratios when local information about the group is available at each transmitter. Existence and uniqueness are proven for Nash plays in games involving automata with continuous sets of moves corresponding to these criteria.

 

Markov Processes over Denumerable Products of Spaces, Describing Large Systems of Automata
L. N. Vasershtein
pp. 47–52

Abstract—A criterion is obtained for the uniqueness of the stationary probabilities of a Markov operator. This criterion is useful for operators related to homogeneous automata games. Similar methods are applicable to certain problems in statistical physics, biology, etc.

 

Propagation of Errors in the Decoding of Uniform Convolutional Codes
A. L. Larin and E. M. Gabidulin
pp. 53–57

Abstract—The effect of error propagation for certain convolutional codes is investigated on the basis of the concept of the stability of the decoder. Among the PO-codes of J. Massey [Threshold Decoding, Cambridge, MIT Press, 1963] we select those in which the effect of decoding errors is extended to an arbitrary distance. For a class of uniform convolutional codes we find the length of the region in which decoding errors have an effect.

 

Optimum Power-Sensitive Reception in the Optical Band
A. S. Drikker and R. R. Krasovskii
pp. 57–60

Abstract—The structure of the optimum receiver is determined by calculating the probability ratio. We take as the initial expression which describes the photocurrent distribution, a negative binomial distribution obtained on the assumption that the power fluctuations of the radiation received are described by a gamma distribution and that the radiation is polarized. Approximations for the cases of strong and weak signals are considered. The detection characteristics are constructed.

 

One Machine Recognition Problem in Self-training
B. A. Golovkin
pp. 60–62

Abstract—Machine recognition in self-training mode is formulated as a combinatorial-type extremal problem. It is shown that an exact solution can be obtained by integer linear programming; approximate-solution algorithms are indicated.

 

Alternative Choice of Structure in One Elementary Communications Network
V. V. Kiryukhin
pp. 62–64

Abstract—Optimization of a centralized communications network (a combination of radial and linear structures) is considered. Minimization of the average number of peripheral points cut off from the center is the optimality criterion. It is shown that there is no combination network in the class of structures optimal by this criterion: only a linear or a radial structure can be optimal.

 

Asymptotic Investigation of Stationary State Probability Distribution for One Class of Single-Line Queueing Systems (Without Memory)
V. A. Ivnitskii
pp. 64–71

Abstract—A class of single-line queueing systems with Markovian input process is considered. The amount of work performed for a demand suffering interrupted service is not stored. The stationary state probability distribution is analyzed asymptotically for this class of systems with low-intensity input process.