Discrete time adaptive impulsive control theory (Q1821082)

This paper consists of two parts. The first part investigates discrete- time impulsive control with a long run average cost criterion. Under suitable assumptions, the author characterizes an \(\epsilon\)-optimal impulsive strategy and, in the case of a Feller process, an optimal strategy is given. In the second part adaptive impulse controls are considered. Using the modified maximum likelihood estimator he estimates the unknown parameter and, appealing to the results of the first part, an optimal control is constructed. Moreover, the relation between optimal controls with known true parameter and optimal adaptive controls is given.