Recent Advances in Reinforcement Learning

Vidyasagar, Mathukumalli (2020) Recent Advances in Reinforcement Learning. In: 2020 American Control Conference, ACC 2020, 1-3 July 2020, Denver.

Text
Proceedings_of_the_American_Control_Conference.pdf - Published Version
Available under License Creative Commons Attribution.
Download (284kB)

Abstract

In this paper, we give a brief review of Markov Decision Processes (MDPs), and how Reinforcement Learning (RL) can be viewed as MDP where the parameters are unknown. Specific topics discussed include the Bellman equation and the Bellman operator, and value and policy iterations for MDPs, together with recent "empirical" approaches to solving the Bellman equation and applying the Bellman iteration. In addition to the well-established method of Q-learning, we also discuss the more recent approach known as Zap Q-learning. © 2020 AACC.

[error in script]

IITH Creators:

IITH Creators	ORCiD
Vidyasagar, Mathukumalli	https://orcid.org/0000-0003-1057-1942

Item Type:

Conference or Workshop Item (Paper)

Additional Information:

M. Vidyasagar is a National Science Chair, and is with the Indian Institute of Technology Hyderabad, India, This research is supported by the Science and Engineering Research Board (SERB), Government of India. Email: m.vidyasagar@iith.ac.in.

Uncontrolled Keywords:

Bellman equations; Markov Decision Processes; Policy iteration; Q-learning

Subjects:

Electrical Engineering

Divisions:

Department of Electrical Engineering

Depositing User:

. LibTrainee 2021

Date Deposited:

23 Nov 2022 09:11

Last Modified:

23 Nov 2022 09:11