Assistance Required for running Parallel Computing in RL Designer Toolbox in MATLAB

When attempting to enable parallel computation, I receive the following error message:

"Unable to open the requested feature. Check your internet connection and proxy settings in MATLAB Web preferences and then try starting the feature again. Detailed information: Error code: -108 Error message: ERR_ADDRESS_INVALID."

I have checked my internet connection and proxy settings as suggested but the issue persists. I have attached a screenshot of the error message for your reference.

Additionally, I am seeking guidance on the proper configuration of the files to be added in parallel computation toolbox, mySetup and myCleanup functions required for parallel computing. I have included a sample of my current mySetup and myCleanup functions below for your review:

matlab

Copy code

function mySetup()
load('RL_input_data.mat', 'RL_input_data');

% Define and check the script file to run
scriptFile = 'main.mlx';
% Run the script file
run(scriptFile);
end

function myCleanup()
% Save workspace or final state
% save('workerLastState.mat');

% Clean up global variables or close files
% clear global sharedData

% Additional cleanup tasks
end

I would appreciate any recommendations or examples of mySetup and myCleanup functions that would be more suitable for my scenario.

Please find attached the images of my MATLAB configuration settings and the error message for your reference.

Thank you for your support.

Best regards,

Manikanta Polepally

[Walter Roberson]

Images of the configuration setting and error message are currently missing.

[Walter Roberson]

I am not sure, but it sounds as if your system has a firewall that is preventing communication between processes on the same host.

Answer 1

[Kautuk Raj]

I see that you are encountering an error when attempting to enable parallel computation in the Reinforcement Learning Toolbox in MATLAB, which suggests checking your internet connection and proxy settings. Despite verifying these settings, the issue persists.

I suspect you might be trying to carry out parallel training with LSTM-based models and multi-agent environments, which is not supported in the Reinforcement Learning Toolbox. More about this can be found at this MATLAB Answers post: https://www.mathworks.com/matlabcentral/answers/894942-reinforcement-learning-with-parallel-computing

I trust this explanation will greatly assist you in addressing your query.