This book does not intend to be a comprehensive text on stochastic approximation (SA) but rather a “bite-sized” one. Besides the basic convergence analysis of SA procedures, attention is paid to topics connected with their implementation, such as multiple timescales or asynchronous schemes. A short chapter is devoted to the constant stepsize algorithms. Methodologically, the ODE approach has been adopted, which treats the SA algorithm as a noisy discretization of an ordinary differential equation. An overview of several applications is given. Some less standard mathematical requirements can be found in the appendices.
Reviewer:
vdup