Linear Programming and Constrained Average Optimality for General Continuous-Time Markov Decision Processes in History-Dependent Policies.
Xianping GuoYonghui HuangXinYuan SongPublished in: SIAM J. Control. Optim. (2012)
Keyphrases
- average cost
- markov decision processes
- stationary policies
- linear programming
- linear program
- optimal policy
- dynamic programming
- state space
- finite state
- optimal control
- action sets
- policy iteration
- reinforcement learning
- finite horizon
- markov decision process
- special case
- long run
- markov decision problems
- average reward
- discounted reward
- reachability analysis
- markov chain
- np hard
- planning under uncertainty
- decision theoretic planning
- infinite horizon
- finite number
- initial state
- decision processes
- decision problems
- multistage
- optimal solution
- objective function
- reinforcement learning algorithms
- lot sizing
- risk sensitive
- state abstraction
- factored mdps
- state and action spaces
- decentralized control
- computational complexity
- multi agent