penaltykick

Batch Reinforcement Learning in a Complex Domain

Shivaram Kalyanakrishnan and Peter Stone. Batch Reinforcement Learning in a Complex Domain. In The Sixth International Joint Conference on Autonomous Agents and Multiagent Systems, pp. 650–657, ACM, New York, NY, USA, May 2007.
BEST PAPER AWARD NOMINEE.
AAMAS-2007

Download

[PDF]186.4kB  [postscript]384.8kB  

Abstract

Temporal difference reinforcement learning algorithms are perfectly suited to autonomous agents because they learn directly from an agent's experience based on sequential actions in the environment. However, their most common algorithmic variants are relatively inefficient in their use of experience data, which in many agent-based settings can be scarce. In particular, they make just one learning ``update'' for each atomic experience. Batch reinforcement learning algorithms, on the other hand, aim to achieve greater data efficiency by saving experience data and using it in aggregate to make updates to the learned policy. Their success has been demonstrated in the past on simple domains like grid worlds and low-dimensional control applications like pole balancing. In this paper, we compare and contrast batch reinforcement learning algorithms with on-line algorithms based on their empirical performance in a complex, continuous, noisy, multiagent domain, namely RoboCup soccer Keepaway. We find that the two batch methods we consider, Experience Replay and Fitted Q Iteration, both yield significant gains in sample complexity, while achieving high asymptotic performance.

BibTeX Entry

@InProceedings{AAMAS07-kalyanakrishnan,
author="Shivaram Kalyanakrishnan and Peter Stone",
title="Batch Reinforcement Learning in a Complex Domain",
booktitle="The Sixth International Joint Conference on Autonomous Agents and Multiagent Systems",
month="May",year="2007",
isbn = "978-81-904262-7-5",
pages = "650--657",
location = "Honolulu, Hawaii",
publisher = "ACM",
address = "New York, NY, USA",
abstract={Temporal difference reinforcement learning algorithms are
perfectly suited to autonomous agents because they learn directly from
an agent's experience based on sequential actions in the environment.
However, their most common algorithmic variants are relatively
inefficient in their use of experience data, which in many agent-based
settings can be scarce. In particular, they make just one learning
``update'' for each atomic experience. Batch reinforcement learning
algorithms, on the other hand, aim to achieve greater data efficiency
by saving experience data and using it in aggregate to make updates to
the learned policy. Their success has been demonstrated in the past on
simple domains like grid worlds and low-dimensional control
applications like pole balancing. In this paper, we compare and
contrast batch reinforcement learning algorithms with on-line
algorithms based on their empirical performance in a complex,
continuous, noisy, multiagent domain, namely RoboCup soccer Keepaway.
We find that the two batch methods we consider, Experience Replay and
Fitted Q Iteration, both yield significant gains in sample complexity,
while achieving high asymptotic performance.},
}

Valid CSS!
Valid XHTML 1.0!