Constrained Optimization for Average Cost Continuous-Time Markov Decision Processes
Xianping Guo · IEEE Transactions on Automatic Control · 2007
To the best of our knowledge, this note is a first study on a multiconstrained optimization problem for average cost continuous-time Markov decision processes in polish spaces. The optimality criterion is the long-run expected average cost, and the constraints are imposed on similar average costs. We give suitable conditions under which the existence of a constrained optimal stationary policy is shown by introducing occupation measures and using a martingale technique. Also, we illustrate our conditions with controlled Potlach processes.