An Adaptive Sampling Algorithm for Policy Evaluation

Joseph, AG and Bhatnagar, S (2019) An Adaptive Sampling Algorithm for Policy Evaluation. In: 5th Indian Control Conference, ICC 2019, 9 January 2019- 11 January 2019, Delhi, pp. 2-7.

PDF ICC _2019.pdf - Published Version Restricted to Registered users only Download (15MB) | Request a copy

Abstract

In this paper, we propose two efficient and stable adaptive sampling algorithms for policy evaluation in reinforcement learning under linear function approximation. The computational complexities of the algorithms scale quadratically and linearly on the number of features respectively. The empirical analysis shows that the algorithms converge to the neighbourhood of the fixed point of the projected Bellman equation faster than the respective state-of-the-art algorithms.

Item Type:	Conference Paper
Publication:	2019 5th Indian Control Conference, ICC 2019 - Proceedings
Publisher:	Institute of Electrical and Electronics Engineers Inc.
Additional Information:	The copyright for this article belongs to Institute of Electrical and Electronics Engineers Inc.
Keywords:	Dynamic programming; Learning algorithms; Reinforcement learning, Adaptive sampling algorithms; Bellman equations; Empirical analysis; Fixed points; Linear functions; Neighbourhood; Policy evaluation; State-of-the-art algorithms, Approximation algorithms
Department/Centre:	Division of Electrical Sciences > Computer Science & Automation
Date Deposited:	27 Dec 2022 05:22
Last Modified:	27 Dec 2022 05:22
URI:	https://eprints.iisc.ac.in/id/eprint/78571

Actions (login required)

View Item