notesum.ai

Published at December 9

Edge Delayed Deep Deterministic Policy Gradient: efficient continuous control for edge scenarios

cs.LG

cs.AI

Released Date: December 9, 2024

Authors: Alberto Sinigaglia¹, Niccolò Turcato², Ruggero Carli², Gian Antonio Susto²

Aff.: ¹Human-Inspired Technology Research Center, University of Padova; ²Department of Information Engineering, University of Padova

Arxiv: http://arxiv.org/pdf/2412.06390v1

Refer to caption

Algorithm	Time
EdgeD3	$214.0\pm 7.1$ ms
DDPG	$285.5\pm 7.4$ ms ( $-25.0\%$ )
TD3	$308.2\pm 2.7$ ms ( $-30.5\%$ )
SAC^d	$320.90\pm 3.6$ ms ( $-33.3\%$ )
SAC^o	$492.91\pm 2.9$ ms ( $-56.8\%$ )