Abstract:
The one-sided convergence of the modified Robbins–Monro process is studied [1]. A method for choosing the algorithm constants is designed such that the probability for not exceeding the unknown root of the regression equation is arbitrarily close to 1.
Presented by the member of Editorial Board:A. V. Nazin
Citation:
T. P. Krasulina, Yu. O. Yatel, “A probability estimate for not exceeding the unknown threshold by the Robbins–Monro algorithm”, Avtomat. i Telemekh., 2005, no. 3, 91–96; Autom. Remote Control, 66:3 (2005), 422–426