Bad-Policy Density: A Measure of Reinforcement Learning Hardness.
David AbelCameron AllenDilip ArumugamD. Ellis HershkowitzMichael L. LittmanLawson L. S. WongPublished in: CoRR (2021)
Keyphrases
- reinforcement learning
- optimal policy
- policy search
- similarity measure
- action selection
- markov decision process
- markov decision processes
- actor critic
- action space
- reinforcement learning algorithms
- reward function
- function approximation
- learning algorithm
- policy gradient
- control policy
- partially observable
- state space
- policy iteration
- phase transition
- distance measure
- reinforcement learning problems
- computational complexity
- approximate dynamic programming
- partially observable domains
- state and action spaces
- learning process
- markov decision problems
- state action
- partially observable environments
- neural network
- search space
- control policies
- np hard
- dynamic programming
- decision problems
- correlation coefficient
- optimal control