Asymptotically optimal controls of hybrid LQG problems: Summary of results. (Q2712238)

This work gives a summary of the authors' results about control of a linear stochastic system with quadratic cost function for the case where system matrices depend on Markov parameter. The work treats the situation where the state space of the modulating chain is large, but is decomposable into view groups such that the chain fluctuates rapidly within each group and less frequently among groups. To reduce the computational effort, an approximation scheme is presented which yields a nearly optimal solution of the initial problem. A numerical example demonstrates the approximation procedure.NEWLINENEWLINEFor the entire collection see [Zbl 0958.00050].