A Hybrid Approach for Reinforcement Learning Using Virtual Policy Gradient for Balancing an Inverted Pendulum.

Published in: CoRR (2021)

Keyphrases

policy gradient
inverted pendulum
reinforcement learning
actor critic
feedback control
optimal control
function approximation
reinforcement learning algorithms
intelligent control

simulation study
control algorithm
nonlinear systems
fuzzy controller
gradient method
control problems
partially observable markov decision processes
temporal difference
initial conditions

fuzzy systems
function approximators
reinforcement learning methods
mobile robot
adaptive control
model free
input output
state action

approximation methods
fuzzy logic
real time
policy iteration
optimal policy
state space
variance reduction
learning algorithm