Control Frequency Adaptation via Action Persistence in Batch Reinforcement Learning

Alberto Maria Metelli, Flavio Mazzolini, Lorenzo Bisi, Luca Sabbioni, Marcello Restelli
Proceedings of the 37th International Conference on Machine Learning, PMLR 119:6862-6873, 2020.

Abstract

The choice of the control frequency of a system has a relevant impact on the ability of reinforcement learning algorithms to learn a highly performing policy. In this paper, we introduce the notion of action persistence that consists in the repetition of an action for a fixed number of decision steps, having the effect of modifying the control frequency. We start analyzing how action persistence affects the performance of the optimal policy, and then we present a novel algorithm, Persistent Fitted Q-Iteration (PFQI), that extends FQI, with the goal of learning the optimal value function at a given persistence. After having provided a theoretical study of PFQI and a heuristic approach to identify the optimal persistence, we present an experimental campaign on benchmark domains to show the advantages of action persistence and proving the effectiveness of our persistence selection method.

Cite this Paper


BibTeX
@InProceedings{pmlr-v119-metelli20a, title = {Control Frequency Adaptation via Action Persistence in Batch Reinforcement Learning}, author = {Metelli, Alberto Maria and Mazzolini, Flavio and Bisi, Lorenzo and Sabbioni, Luca and Restelli, Marcello}, booktitle = {Proceedings of the 37th International Conference on Machine Learning}, pages = {6862--6873}, year = {2020}, editor = {III, Hal Daumé and Singh, Aarti}, volume = {119}, series = {Proceedings of Machine Learning Research}, month = {13--18 Jul}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v119/metelli20a/metelli20a.pdf}, url = {http://proceedings.mlr.press/v119/metelli20a.html}, abstract = {The choice of the control frequency of a system has a relevant impact on the ability of reinforcement learning algorithms to learn a highly performing policy. In this paper, we introduce the notion of action persistence that consists in the repetition of an action for a fixed number of decision steps, having the effect of modifying the control frequency. We start analyzing how action persistence affects the performance of the optimal policy, and then we present a novel algorithm, Persistent Fitted Q-Iteration (PFQI), that extends FQI, with the goal of learning the optimal value function at a given persistence. After having provided a theoretical study of PFQI and a heuristic approach to identify the optimal persistence, we present an experimental campaign on benchmark domains to show the advantages of action persistence and proving the effectiveness of our persistence selection method.} }
Endnote
%0 Conference Paper %T Control Frequency Adaptation via Action Persistence in Batch Reinforcement Learning %A Alberto Maria Metelli %A Flavio Mazzolini %A Lorenzo Bisi %A Luca Sabbioni %A Marcello Restelli %B Proceedings of the 37th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2020 %E Hal Daumé III %E Aarti Singh %F pmlr-v119-metelli20a %I PMLR %P 6862--6873 %U http://proceedings.mlr.press/v119/metelli20a.html %V 119 %X The choice of the control frequency of a system has a relevant impact on the ability of reinforcement learning algorithms to learn a highly performing policy. In this paper, we introduce the notion of action persistence that consists in the repetition of an action for a fixed number of decision steps, having the effect of modifying the control frequency. We start analyzing how action persistence affects the performance of the optimal policy, and then we present a novel algorithm, Persistent Fitted Q-Iteration (PFQI), that extends FQI, with the goal of learning the optimal value function at a given persistence. After having provided a theoretical study of PFQI and a heuristic approach to identify the optimal persistence, we present an experimental campaign on benchmark domains to show the advantages of action persistence and proving the effectiveness of our persistence selection method.
APA
Metelli, A.M., Mazzolini, F., Bisi, L., Sabbioni, L. & Restelli, M.. (2020). Control Frequency Adaptation via Action Persistence in Batch Reinforcement Learning. Proceedings of the 37th International Conference on Machine Learning, in Proceedings of Machine Learning Research 119:6862-6873 Available from http://proceedings.mlr.press/v119/metelli20a.html.

Related Material