Dissipativity in infinite horizon optimal control and dynamic programming (Q6118398)

New types of optimal control discrete-time problems with both finite and infinite horizon are studied. The finite horizon optimal control problem is formulated as follows: \[ J_{\tau}^{\psi}(x(.),y(.)) \equiv \Sigma_{t=0}^{\tau-1} h(x(t),u(t) + \psi(\tau) ~\rightarrow \min \] subject to \[ x(0) = x,~ x(t+1) = f(x(t), u(t)),~(x(t),u(t)) \in Z,~ t \in T = \{0, \dots, \tau - 1\},~x(\tau) \in X, \] where \(x(t) \in X \subset R^n\), \(X\) is compact, \(f: Z ~\rightarrow~X\) is continuous, \(\psi: X \rightarrow R\) is a continuous terminal cost function. The paper uses the dynamic programming technique and investigates possibilities how to select the terminal cost function to be able to formulate the problem as an optimal control problem with infinite horizon. Theoretical results are illustrated by numerical examples. Further research is outlined in the concluding part of the paper