TY - GEN
T1 - Application of Newton's method to action selection in continuous state- and action-space reinforcement learning
AU - Nichols, Barry D.
AU - Dracopoulos, Dimitris C.
PY - 2014
Y1 - 2014
N2 - An algorithm based on Newton's Method is proposed for action selection in continuous state- and action-space reinforcement learning without a policy network or discretization. The proposed method is validated on two benchmark problems: Cart-Pole and double Cart-Pole on which the proposed method achieves comparable or improved performance with less parameters to tune and in less training episodes than CACLA, which has previously been shown to outperform many other continuous state- and action-space reinforcement learning algorithms.
AB - An algorithm based on Newton's Method is proposed for action selection in continuous state- and action-space reinforcement learning without a policy network or discretization. The proposed method is validated on two benchmark problems: Cart-Pole and double Cart-Pole on which the proposed method achieves comparable or improved performance with less parameters to tune and in less training episodes than CACLA, which has previously been shown to outperform many other continuous state- and action-space reinforcement learning algorithms.
UR - http://www.scopus.com/inward/record.url?scp=84962022092&partnerID=8YFLogxK
M3 - Conference contribution
AN - SCOPUS:84962022092
T3 - 22nd European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning, ESANN 2014 - Proceedings
SP - 141
EP - 146
BT - 22nd European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning, ESANN 2014 - Proceedings
PB - i6doc.com publication
T2 - 22nd European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning, ESANN 2014
Y2 - 23 April 2014 through 25 April 2014
ER -