Stochastic algorithms of optimization with Markov noises in gradient measurements

Convergence is investigated and rates of convergence are comparatively analyzed for gradient stochastic procedures in the presence of additive noises described by difference equations.