Weighted Policy Constraints for Offline Reinforcement Learning.

Zhiyong Peng Changlin Han Yadong Liu Zongtan Zhou

Published in: AAAI (2023)

Keyphrases

reinforcement learning
optimal policy
policy search
state space
markov decision process
partially observable
machine learning
function approximation

reward function
action selection
linear constraints
state and action spaces
reinforcement learning algorithms
real time
markov decision processes
policy iteration

model free
policy evaluation
markov decision problems
state action
partially observable domains
multi agent
reinforcement learning problems
semi supervised

approximate dynamic programming
policy gradient
control policies
control policy
function approximators
decision problems
constraint programming
infinite horizon