Adaptive Systems

Dynamic stochastic approximation of nonstationary solution to a conditional extremum problem

V. Ya. Katkovnik, V. E. Kheisin

Leningrad

Abstract: A universal misalignment function is proposed whose application in conventional unconditional optimization algorithms leads to algorithms which solve the conditional extremum problem rather than to a gradient. That function leads to algorithms of dynamic stochastic approximation.

UDC: 62-506


Received: 18.07.1979

 English version: Automation and Remote Control, 1980, 41:6, 794–801

Bibliographic databases:

• 
•