Policy Transfer using Reward Shaping.
Tim BrysAnna HarutyunyanMatthew E. TaylorAnn NowéPublished in: AAMAS (2015)
Keyphrases
- reward shaping
- markov decision problems
- reinforcement learning
- optimal policy
- reward function
- markov decision process
- complex domains
- state space
- policy search
- reinforcement learning algorithms
- transfer learning
- action selection
- markov decision processes
- linear programming
- decision processes
- transition probabilities
- partially observable
- case based reasoning
- multi agent
- policy gradient
- transition model
- optimal solution
- decision making