Scholarlink

In this paper, we explore the optimization of hyperparameters for the Soft Actor-Critic (SAC) and Proximal Policy Optimization (PPO) algorithms using the Tree-structured Parzen Estimator (TPE) in the context of robotic arm control with seven Degrees of Freedom (DOF). Our results demonstrate a significant enhancement in algorithm performance, TPE improves the success rate of SAC by 10.48% points and PPO by 34.28% points, where models trained for 50K episodes. Furthermore, TPE enables PPO to converge to a reward within 95% of the maximum reward 76% faster than without TPE, which translates to about 40K fewer episodes of training required for optimal performance. Also, this improvement for SAC is 80% faster than without TPE.  This study underscores the impact of advanced hyperparameter optimization on the efficiency and success of deep reinforcement learning algorithms in complex robotic tasks.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Optimizing Deep Reinforcement Learning for Adaptive Robotic Arm Control

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Jonaid Shianifar,
  • Michael Schukat,
  • Karl Mason

摘要

In this paper, we explore the optimization of hyperparameters for the Soft Actor-Critic (SAC) and Proximal Policy Optimization (PPO) algorithms using the Tree-structured Parzen Estimator (TPE) in the context of robotic arm control with seven Degrees of Freedom (DOF). Our results demonstrate a significant enhancement in algorithm performance, TPE improves the success rate of SAC by 10.48% points and PPO by 34.28% points, where models trained for 50K episodes. Furthermore, TPE enables PPO to converge to a reward within 95% of the maximum reward 76% faster than without TPE, which translates to about 40K fewer episodes of training required for optimal performance. Also, this improvement for SAC is 80% faster than without TPE.  This study underscores the impact of advanced hyperparameter optimization on the efficiency and success of deep reinforcement learning algorithms in complex robotic tasks.