PROBLEMS OF INFORMATION TRANSMISSION
A translation of Problemy Peredachi Informatsii
Information Transfer by a Quantum
Electromagnetic Field in an Absorbing Medium
V. D. Vainshtein
pp. 110
AbstractA quantum electromagnetic information-transmission channel is investigated with regard for the absorptive and emissive properties of the medium in which the signal propagates. It is shown that absorption and induced emission processes can influence the channel rate significantly in the optical range.
Universal Decoding for Symmetric Channels
V. D. Goppa
pp. 1118
AbstractA new decoding scheme is proposed for linear codes in symmetric channels.
Estimation of the Error-Correction
Complexity for Gallager Low-Density Codes
V. V. Zyablov and M. S. Pinsker
pp. 1828
AbstractIt is shown that among the Gallager low-density binary codes there are codes with decoding for the correction of all errors up to multiplicity $\alpha n$ and with decoding complexity of order $n\log n$, where $n$ is the code length and $\alpha$ is a positive number.
Optimal Encoding of a Source on the Basis of Observations
R. E. Krichevskii
pp. 2932
AbstractThe optimal encoding problem is discussed for a source whose distribution function is not known beforehand and is estimated from observations thereof.
On BahlChien Burst-Error-Correcting Codes
G. L. Tauglich and G. M. Tenengol'ts
pp. 3343
AbstractNew bounds are derived for the length of single and double burst errors corrected by BahlChien codes. Simple decoding schemes are proposed for these codes.
Nonbinary Arithmetic Codes with Large Minimum Distance
I. M. Boyarinov
pp. 4449
Abstract$AN$ codes over an arbitrary base $r$ are discussed. A unique canonical $r$-ary representation having a minimum number of nonzero terms is found for the number $N$. An expression is obtained for the minimum distance of $r$-ary $AN$ codes with large distance for two metrics: arithmetic and Hamming.
Discriminant Analysis of Normal
Populations Represented by Large Samples
L. G. Malinovskii
pp. 5056
AbstractAn algorithm is given for the plotting of all extremal values of various criteria of discriminant directions. It is shown that this algorithm can be used in the case of large samples to test a hypothesis concerning the properties of the probability distribution functions and to formulate classification rules.
Ergodic Theorem for a Class of Markov
Processes with Local Interaction
R. K. Romanovskii
pp. 5762
AbstractA necessary and sufficient condition is found for the existence of a steady state for certain locally interacting Markov process. This condition is formulated in terms of the joint spectrum of $n$ stochastic matrices controlling the process. In the case $n=1$ it coincides with the ergodic theorem for a homogeneous Markov chain.
Estimation of the Blocking Probability
for Switching Circuits by Means of Probability Graphs
I. V. Koverninskii
pp. 6371
AbstractVarious probability graph models are described. They are used to obtain asymptotic estimates for the blocking probability of a particular class of cascaded switching circuits. The results of a computer simulation of the probability graphs are also given. These results are used to assess the possibility of using the given models to describe the operation of switching circuits.
Control of a Point Process of
Observable Tokens Associated with the Detection of a Point Process of Objects
and Measurement of Their Coordinates
Yu. V. Zhulina
pp. 7280
AbstractThe control of a point process of observable tokens generated by a point process of objects or events plus background noise is discussed. An algorithm is proposed and analyzed for the control of the token process on the basis of the criterion of maximum quantity of information about the object process.
Mappings of Pattern Space in Statistical Decision Theory
T. Farago
pp. 8084
AbstractCertain estimates for the quality of a mapping of pattern space (observation space) from the standpoint of subsequent decisions as to the membership of a pattern in one of two classes.
Efficient Encoding Under Varying Source Statistics
Yu. M. Baranov
pp. 8587
AbstractIt is shown that for source statistics varying over a wide range the Shannon encoding algorithm with constant parameters provides a fully acceptable compression of information for many applications.
Additive Group Codes for Defect Correction
B. S. Tsybakov
pp. 8890
AbstractGroup additive codes for the correction of defects are described. These codes are asymptotically worse than optimal nonlinear codes, but they admit simpler implementation of the encoding and decoding operations.
Technique for the Construction of Invariant Statistics
V. A. Bogdanovich
pp. 9193
AbstractA technique id discussed for the construction of a maximum invariant, in terms of which is expressed an invariant rule for the testing of composite statistical hypothesis. The technique can be used for the solution of signal-detection problems under the conditions of parametric a priori uncertainty about the distribution functions of observed data. Examples are given.
Nomograms for Calculation of the
Average Waiting Time for Priority Systems
V. M. Kozub and V. V. Chibisov
pp. 9496
AbstractNomograms are proposed for calculation of the average waiting time for calls in a single-server queuing system with preemptive and nonpreemptive priorities, Poisson input, and exponential or constant service time. Numerical examples and tables are given to illustrate the sequence of operations for the given nomograms.