Dynamic parameters in Sequential Decision Making

Sequential Decision Making (SDM) problems optimize over the sequence of actions (or, decisions) taken to minimize underlying cumulative cost. These are referred as policy SDM. Often these comprise additional (fixed and manipulable) parameters; objective is determine optimal well manipulable parameters that minimizes SDM In this paper we address class characterized by dynamic where dynamics pre-specified for a subset others. The parameter time-varying such associated cost gets minimized at each time instant. To end, develop control-theoretic framework design it tracks values parameters, simultaneously determines policy. Our methodology builds upon Maximum Entropy Principle (MEP) based addresses SDMs. More precisely, above results into smooth approximation which utilize control Lyapunov function. We show under resulting law asymptotically track local optimal, proposed Lipschitz continuous bounded, given set values. simulations demonstrate efficacy our methodology.