Stochastic approximation and its applications (Q1851214): Difference between revisions

Estimating unknown parameters based on observation data is the basic problem for stochastic modeling. Some of the methods of estimation can be transformed into methods for searching for zeroes \(x^\circ\) of an unknown function \(f(\cdot): f(x^\circ)=0\). \textit{H. Robbins} and \textit{S. Monro} [Ann. Math. Stat. 22, 400-407 (1951; Zbl 0054.05901)] developed an algorithm or stochastic approximation method for solving such problems. The present author introduces a modified technique termed as expanding truncation technique combined with the trajectory-subsequence (TS) method to study problems of estimation via stochastic approximation. The contents of the book are as follows: Chapter 1: Robbins-Monro algorithm; Chapter 2: Stochastic approximation algorithms with expanding truncations; Chapter 3: Asymptotic properties of stochastic approximation algorithms; Chapter 4: Optimization by stochastic approximation; Chapter 5: Application to signal processing; Chapter 6: Application to systems and control. The book is written in theorem/proof format and should be of interest to those who work in the theory of stochastic approximation and related applications.