Linear Quadratic Control Problem For Jump Linear Systems With No Observation Of The Markov Chain States

The subject matter of this paper is the study of a stochastic control problem for a class of linear systems subject to Markovian jumps among different forms and quadratic cost. It is assumed that the system is partially observable in the sense that we do not have access to the jumping parameters and the control can only depend on the present value of the linear state variable. The main feature of the approach here is that we do not recast the problem as one with complete observations, and the solution is determined by a set of interconnected Riccati equations, similar to the complete observation case. A peculiar attribute of the approach here is a robust flavor and the explicit form for the optimal control policy.21392139