Market-Based Reinforcement Learning in Partially Observable Worlds

Ivo Kwee; Marcus Hutter; Juergen Schmidhuber

Market-Based Reinforcement Learning in Partially Observable Worlds

Artificial Intelligence 2007-05-23 v1 Machine Learning Multiagent Systems Neural and Evolutionary Computing

Authors: Ivo Kwee , Marcus Hutter , Juergen Schmidhuber

Abstract

Unlike traditional reinforcement learning (RL), market-based RL is in principle applicable to worlds described by partially observable Markov Decision Processes (POMDPs), where an agent needs to learn short-term memories of relevant previous events in order to execute optimal actions. Most previous work, however, has focused on reactive settings (MDPs) instead of POMDPs. Here we reimplement a recent approach to market-based RL and for the first time evaluate it in a toy POMDP setting.

Keywords

reinforcement learning markov decision processes multi-agent reinforcement learning

Cite

@article{arxiv.cs/0105025,
  title  = {Market-Based Reinforcement Learning in Partially Observable Worlds},
  author = {Ivo Kwee and Marcus Hutter and Juergen Schmidhuber},
  journal= {arXiv preprint arXiv:cs/0105025},
  year   = {2007}
}

Comments

8 LaTeX pages, 2 postscript figures

Related papers

View all related →