List your company
ResearchFormerly WizWand

Reinforcement Learning on Inverted Double Pendulum

9,359.92Avg Episode RewardSAC
-291.77922,213.95044,719.687,225.4096Nov 2, 2023Apr 3, 2024Sep 3, 2024Feb 3, 2025Jul 6, 2025Dec 6, 2025May 8, 2026
Updated 5mo ago

Evaluation results

MethodLinks
SAC2023.119,359.92
ESPL2023.119,359.9
A2C2023.119,359.81
TD32023.119,359.25
ACKTR2023.119,359.06
PPO2023.119,356.59
DDPG2023.119,347.1
TRPO2023.119,188.43
DSP2023.119,149.9
Regression2023.11637.2
R2PONumber of independent runs=102026.05254.04
R2PONumber of runs=10, Aggregation protocol=averaged across all training iterations2026.05158.51
ProPS+Number of independent runs=102026.05128.81
ProPSNumber of independent runs=102026.05112.18
TRPOSource=Best SB3, Number of independent runs=102026.0598.25
ProPS+Number of runs=10, Aggregation protocol=averaged across all training iterations2026.0586.71
TRPONumber of runs=10, Aggregation protocol=averaged across all training iterations, Source=Best SB32026.0586.04
ProPSNumber of runs=10, Aggregation protocol=averaged across all training iterations2026.0579.44