Simplex Algorithm for Countable-State Discounted Markov Decision Processes.
Ilbin LeeMarina A. EpelmanH. Edwin RomeijnRobert L. SmithPublished in: Oper. Res. (2017)
Keyphrases
- markov decision processes
- state space
- simplex algorithm
- average cost
- optimal policy
- action space
- partially observable
- finite state
- discounted reward
- infinite horizon
- dynamic programming
- markov chain
- real time dynamic programming
- discount factor
- reinforcement learning
- total reward
- markov decision process
- finite horizon
- decision theoretic planning
- average reward
- linear programming
- transition matrices
- policy iteration
- heuristic search
- primal dual
- stationary policies
- long run
- collaborative filtering
- linear programming problems
- interior point methods
- state and action spaces
- feasible solution
- lower bound
- optimal control
- bayesian networks