Characterization of the Optimal Policy for a Multidimensional Parabolic Singular Stochastic Control Problem.
Marcin BorycLukasz KrukPublished in: SIAM J. Control. Optim. (2016)
Keyphrases
- stochastic control
- optimal policy
- control problems
- optimal control
- reinforcement learning
- queueing systems
- dynamic programming
- infinite horizon
- brownian motion
- decision problems
- markov decision processes
- operations management
- state space
- finite horizon
- inventory level
- finite state
- long run
- multistage
- state dependent
- markov decision process
- average reward
- initial state
- sufficient conditions
- policy iteration
- rl algorithms
- queueing networks
- average cost
- markov processes
- queue length
- function approximation