Abstract
On reinforcement learning researches, even though environments have continuous state space, many RL algorithms are assumed to be on a discrete state space. Typically, most approaches which treat continuous state and action spaces, just discretise these spaces. In this paper, to treat the continuous state space, we propose a RL algorithm which based on the locally weighted regression.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have