Information processing in a three-actions dynamic decision model (Q1205685)

The authors introduce a finite-horizon stochastic dynamic program to maximize the expected utility of wealth at the end of the planning horizon. At each stage three actions are admissible. If action 1 is choosen the decision maker is able to learn about the unknown distribution of the stochastic outcomes. In case of action 2 (resp. 3) the learning process is interrupted (resp. stopped). Monotonicity arguments are used to give structural results of an optimal strategy.