Reachability and Differential Based Heuristics for Solving Markov Decision Processes.
Shoubhik DebnathLantao LiuGaurav S. SukhatmePublished in: ISRR (2017)
Keyphrases
- markov decision processes
- state space
- transition matrices
- semi markov decision processes
- heuristic search
- optimal policy
- finite state
- reinforcement learning
- policy iteration
- reachability analysis
- average reward
- factored mdps
- decision theoretic planning
- dynamic programming
- markov decision problems
- planning under uncertainty
- infinite horizon
- partially observable
- average cost
- finite horizon
- stochastic shortest path
- decision processes
- model based reinforcement learning
- reinforcement learning algorithms
- action space
- heuristic function
- state abstraction
- search algorithm
- risk sensitive
- reward function
- search space
- admissible heuristics
- state and action spaces
- markov chain
- decision diagrams
- data mining
- markov decision process
- linear programming