How to log/record signal data from a reinforcement learning (RL) agent/environment over multiple training episodes using Simulink?

1 次查看（过去 30 天）

显示更早的评论

Krupa Prag 2020-10-13

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/612411-how-to-log-record-signal-data-from-a-reinforcement-learning-rl-agent-environment-over-multiple-tra

编辑： Krupa Prag 2020-10-15

采纳的回答： Emmanouil Tzorakoleftherakis

I am training an RL agent in Simulink, as shown here.

I would like to record the data for each episode during the training process.

I have considered the functionality of: 'Log Selected Signals'. However, this does not give/return anything to the workspace.

The environment used is designed using blocks from the powerlib.

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

请先登录，再回答此问题。

采纳的回答

Emmanouil Tzorakoleftherakis 2020-10-15

2
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/612411-how-to-log-record-signal-data-from-a-reinforcement-learning-rl-agent-environment-over-multiple-tra#answer_514953

If you call 'train' with an output argument, you will get many useful statistics from training (see here). Would this work?

1 个评论
显示 -1更早的评论隐藏 -1更早的评论

Krupa Prag 2020-10-15

编辑：Krupa Prag 2020-10-15

Thank you for your response.

请先登录，再进行评论。

更多回答（0 个）

请先登录，再回答此问题。

类别

Simulink

在 Help Center 和 File Exchange 中查找有关 Simulink 的更多信息

标签

产品

版本

R2020a

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Translated by