On Convergence of Discrete Stochastic

The discrete stochastic approximation procedure of Robbins-Monroe type

is considered, where {~i} is a martingale-difference and hi} is nonrandom sequence of positive numbers. Results on the almost-sure asymptotic stability of the procedure and the degree of convergence of lzn 12 ~ 0 are obtained. The approach is based on the general theory of martingales and Lyapunov-Krasovskii functional method.