This is a critical review of recent results in stochastic approximation methods for optimization problems. Part II deals with constrained optimization; here, the stochastic variants of the penalty function method, the method of feasible directions and the method of Lagrange functions are discussed.