In the optimal design and cootrol of preparative chromatographic processes,the obstacles appear when one tries to link the Wilson's framework of chromatographic theories based on partial differential equations(PDEs) with the Eulerian presentation to optimal control approaches based on discrete time states,such as Markov decision processes(MDP) or Model predictive control(MPC).

Based on continuous-time Markov decision processes, a power management policy optimization approach is proposed.

An algorithm to optimize policy in call admission control is presented by using the method of Markov decision processes combined with performance potentials.

An algorithm of reinforcement learning for finite-horizon Markov decision processes;

Based on the theory of Markov decision processes,the policy iteration and online policy iteration algorithms for performance optimization are provided.

Markov Decision Process(MDP)model is the general frame for solving reinforcement learning problems.

For the problem of value iteration(VI) optimization of the compact action set in semi-Markov decision process,a unified standard VI algorithm directly based on the equivalent infinitesimal generator under both discount and average criteria was proposed with the proof of convergence.

Discrete event dynamic systems (DEDSs) are generally existing man-made systems in the real world, and semi-Markov decision process (SMDP) is one of the major models for such systems.