The Dual Approch in an Infinite Horizon Model with a Time-Varying Parameter

In a previous paper Amman and Tucci (2017) discuss the DUAL control method, based on Tse and Bar-Shalom (1973) and (Kendrick, 1981) seminal works, applied to the BMW infinite horizon model with an unknown but constant parameter. In these pages the DUAL solution to the BMW infinite horizon model with one time-varying parameter is reported. The special case where the desired path for the state and control are set equal to 0 and the linear system has no constant is considered. The appropriate Riccati quantities for the augmented system are derived and the time- invariant feedback rule are defined following the same steps as in Amman and Tucci (2017). Finally the new approximate cost-to-go is presented. Two cases are considered. In the first one the optimal control is selected using the updated estimate of the time-varying parameter in the model. In the second one only an old estimate of that parameter is available at the time the decision maker chooses her/his control. For the reader’s sake, most of the technical derivations are confined to a number of short appendices.