File(s) under permanent embargo

Action selection methods in a robotic reinforcement learning scenario

conference contribution
posted on 01.01.2018, 00:00 authored by F Cruz Naranjo, P Wuppen, A Fazrie, C Weber, S Wermter
Reinforcement learning allows an agent to learn a new task while autonomously exploring its environment. For this aim, the agent chooses an action to perform among the available ones for a certain state. Nonetheless, a common problem for a reinforcement learning agent is to find a proper balance between exploration and exploitation of actions in order to achieve an optimal behavior. This paper compares multiple approaches to the exploration/exploitation dilemma in reinforcement learning and, moreover, it implements an exemplary reinforcement learning task within the domain of domestic robotics to show the performance of different exploration policies on it. We perform the domestic task using -greedy, softmax, VDBE, and VDBE-Softmax with online and offline temporal-difference learning. The obtained results show that the agent is able to collect larger and faster reward by using the VDBE-Softmax exploration strategy with both Q-learning and SARSA.

History

Event

Latin American Computational Intelligence Society. Conference (2018 : Gudalajara, Mexico)

Series

Latin American Computational Intelligence Society Conference

Pagination

1 - 6

Publisher

Institute of Electrical and Electronics Engineers

Location

Gudalajara, Mexico

Place of publication

Piscataway, N.J.

Start date

07/11/2018

End date

09/11/2018

ISBN-13

9781538646250

Language

eng

Publication classification

E1.1 Full written paper - refereed

Copyright notice

2019, IEEE

Editor/Contributor(s)

[Unknown]

Title of proceedings

LA-CCI : Proceedings of the 2018 IEEE Latin American Conference on Computational Intelligence