Adaptive policies for discrete-time stochastic control systems with unknown disturbance distribution (Q1099125)

Discrete-time stochastic control systems are considered. The disturbance or driving process is a sequence of independent and identically distributed (i.i.d.) random elements whose common distribution is unknown. The state is completely observable together with realizations of the driving process. Adaptive control policies for the problem of maximizing the discounted reward criterion are introduced and studied. These policies are asymptotically optimal and for each of them uniform approximations of the optimal reward function can be obtained.