PROBLEMS OF INFORMATION TRANSMISSION
A translation of Problemy Peredachi Informatsii

CONTENTS                  

Information and Information Stability of Ergodic Sources
K. Marton
pp. 179–183

Abstract—The information stability of ergodic sources with either discrete time or a continuous input process is proved.

Generalized Bayes Estimates for Transmission in Channels with Feedback
B. Ya. Levit
pp. 184–193

Abstract—A method of transmission in memoryless channels with noiseless feedback is considered, generalizing certain transmission methods described in [J.P.M. Schalkwijk and T. Kailath, IEEE Trans. Inf. Theory, 1966, vol. 12, no. 2, pp. 172–182; J.P.M. Schalkwijk, IEEE Trans. Inf. Theory, 1966, vol. 12, no. 2, pp. 183–189; K.Sh. Zigangirov, Probl. Peredachi Inf., 1967, vol. 3, no. 2, pp. 98–101]. The case in which the average power of the transmitted signals is small is investigated in detail (on the assumption of a certain noise regularity). In this extreme case the proposed transmission method makes it possible to obtain an arbitrarily small error probability for transmission rates whose upper bound coincides asymptotically with the channel carrying capacity.

Majority Decoding of Linear Codes
G. K. Kladov
pp. 194–198

Abstract—An $M$-decoding procedure is described, which generalizes the conventional one-step majority-decoding procedures. It is shown that any linear code correcting an error set closed under multiplication by elements of the basic field can be $M$-decoded. In particular, all binary linear codes can be $M$-decoded.

A Burst-Error-Correcting Code and Its Application for Information Exchange between Computers
G. M. Tenengol'ts, A. A. Davydov, and G. L. Tauglikh
pp. 199–206

Abstract—A single-burst-error-correcting code amenable to simple computer software realization is described. The code has a lower redundancy and a somewhat simpler hardware implementation than the well-known Fire code.

Invariance of Maximum-Likelihood Decisions in the Presence of Nuisance Parameters
V. P. Kuznetsov
pp. 207–214

Abstract—The extraction of a random signal from noise in the presence of nuisance parameters is investigated. Conditions are given under which decisions obtained by the maximum-likelihood principle are invariant under nuisance parameters.

Convergence of Continuous and Discrete Robbins-Monro Procedures in the Case of a Multiple-Root Regression Equation
M. B. Nevel'son
pp. 215–223

Abstract—The limiting behavior of a continuous- or discrete-time stochastic approximation process is investigated in the case of a regression equation having several roots. Subject to certain assumptions, a hypothesis advanced in [V. Fabian, Czech. Math. J., 1960, vol. 10, no. 2, pp. 123–159; Trans. 3rd Prague Conf. on Information Theory, Statistical Decision Functions, and Random Processes, Prague, 1964, pp. 85–105] is proved, namely that unstable points of a deterministic system corresponding to a stochastic approximation process cannot be with positive probability limit points for that process.

Lower Bound for the Power of an Automaton State Code
M. S. Pinsker and Yu. L. Sagalovich
pp. 224–230

Abstract—A lower bound is found for the number $M$ of states of an automaton stable under races and errors of any $t$ or fewer of the total number $n$ of its internal elements. The bound is derived by random encoding of the states of the automaton with code words of length $n$. A set of code words that guarantees the indicated property for an automaton is called an automaton state code. The problem is solved in the general case of $q$-position internal elements, and in this connection two race models are proposed. An upper bound is found for the correcting capacity $t$ of an automaton state code, such that the power $M$ of the code preserves exponential growth. In particular, for $q=2$ the latter is true as long as $t\lt n/16$.

Analysis of Linear Huffman Filters under Arbitrary Initial Conditions
V. D. Parondzhanov
pp. 231–239

Abstract—A mathematical description is formulated for the operation of Huffman filters under any initial conditions in the nonautonomous and autonomous regimes. Methods are given for solving inhomogeneous difference equations over the field $GF(2)$ as well as for calculating the sequence at the output of any section, null sequences, and code rings of type $A$. The dependence of the period of an autonomous sequence on the divisibility of the initial-state polynomial and the truncated characteristic polynomial is determined. The sequential connection of and equivalence conditions for filters are discussed, along with filters of general form, correction of the state vector, etc.

A Stochastic Information-Storage Model
E. A. Ikaunieks
pp. 240–246

Abstract—A stochastic system that consists at time $t$ of $N_t$ automata with $n+1$ states is discussed. The states of the individual automata at time $t+1$ are obtained independently of one another in accordance with probabilities depending on the fractions of the automata in each state at time $t$. It is proved that information on the initial state of the system is retained or lost, depending on rate of growth of $N_t$.

Optimization of Data Flows in Information and Control Systems
D. G. Mikhalev and I. B. Russman
pp. 247–250

Abstract—Certain information and control systems are formalized. Two models are considered, one deterministic and one stochastic. In both cases algorithms are given for the optimum selection of parameters for the systems investigated.

Algorithms for Pattern Recognition Learning under Nonstationary Conditions
Ya. Z. Tsypkin
pp. 251–258

Abstract—The recognition problem in a nonstationary environment is stated, and a solution is given for various information contents, i.e., for various degrees of completeness of the a priori information available on the evolving patterns.

Optimum Measurement of Quantum Variables
V. P. Belavkin and B. A. Grishanin
pp. 259–265

Abstract—A system of equations is derived and analyzed for the determination, in problems involving the estimation of quantum variables, of the form of the optimum measurable (reduced) operators for linear estimation. An approximate solution is given in accordance with the assumption of weak noncommutativity of the optimum nonreduced estimates. It is shown that optimum nonlinear estimation reduces to linear estimation in the Gaussian case.

A Controllable Closed Queueing System
E. B. Veklerov
pp. 266–268

Abstract—The well-known optimum priority rule for systems with Poisson inputs is extended to closed systems under small-traffic conditions.