* Markov chains and their control, general theory, constrained MDPs

This area of research goes back to my Ph.D. thesis (under the guidance of Prof. Adam Shwartz), and I have always remained faithful to it. My main interest is in constrained Markov decision processes, for which the standard dynamic programming techniques are not applicable; the computation of optimal policies requires linear programming or Lagrange techniques. I have been working on different aspects, such as adaptive control, sensitivity analysis, finite-state approximations.