Monotone optimal policies in portfolio liquidation problems.
Daniel CrawfordVikram KrishnamurthyPublished in: ICASSP (2015)
Keyphrases
- optimal policy
- decision problems
- markov decision processes
- state space
- multistage
- infinite horizon
- average reward
- partially observable markov decision processes
- machine learning
- dynamic programming
- np complete
- state dependent
- average reward reinforcement learning
- dynamic programming algorithms
- finite horizon
- partially observable
- lower bound
- reinforcement learning