Approximate Probabilistic Constraints and Risk-Sensitive Optimization Criteria in Markov Decision Processes.
Dmitri A. DolgovEdmund H. DurfeePublished in: AI&M (2004)
Keyphrases
- risk sensitive
- markov decision processes
- optimization criteria
- optimization criterion
- utility function
- optimal control
- reinforcement learning
- finite state
- policy iteration
- optimal policy
- state space
- optimization problems
- dynamic programming
- average cost
- partially observable
- decision processes
- markov decision problems
- reinforcement learning algorithms
- model free
- planning under uncertainty
- average reward
- infinite horizon
- cost function
- expected utility
- machine learning