Adaptive Systems

Non-Markov algorithms for statistical optimization with continuous time. I

O. Yu. Kulchitskii

Leningrad

Abstract: A class of statistical optimization algorithms are described with continuous time and of non-Markov type. The algorithms are specified as a set of differential equations with correlated random disturbances. Sufficient conditions for convergence of these algorithms on the average and with a probability of 1 are given.

UDC: 519.25:518.5


Received: 12.05.1977

 English version: Automation and Remote Control, 1978, 39:5, 680–691

Bibliographic databases:

•