search

jjalcaraz-upct / network-slicing

Network slicing gym environment and a model-based RL agent with kernels
MIT License
36 stars 11 forks source link

how to enable a very large number of PRBs? #2

Closed mhmd97z closed 1 year ago

mhmd97z commented 1 year ago

Hi, I'm running a few experiments using your code. I increased the number of PRBs above 200, but it seems that it is not supported. Could you give some hints on where the bottleneck is and how I can increase the capacity of PRBs? Thanks

jjalcaraz-upct commented 1 year ago

Hi,

Thank you for using my code. The reason why the code doesn’t work for a number of PRBs above 200 is because the files containing the SINR traces were obtained for only 100 PRBs, i.e. they contain 100 rows of SINR samples. Therefore, when the number of PRBs exceeds 100, the code stacks duplicate rows on top of the original ones, so that there is maximum frequency distance between a duplicate and the original. The code can get up to 200 PRBs with all rows duplicated. All this is done in the SINRSelectiveFading class, defined in channel_models.py. It is not hard to modify it to allow for a larger PRB extension.

BR,

Juanjo

From: Mohammad Zangooei @.> Reply-To: jjalcaraz-upct/network-slicing @.> Date: Friday, October 14, 2022 at 3:09 AM To: jjalcaraz-upct/network-slicing @.> Cc: Subscribed @.> Subject: [jjalcaraz-upct/network-slicing] how to enable a very large number of PRBs? (Issue #2)

Hi, I'm running a few experiments using your code. I increased the number of PRBs above 200, but it seems that it is not supported. Could you give some hints on where the bottleneck is and how I can increase the capacity of PRBs? Thanks

— Reply to this email directly, view it on GitHubhttps://github.com/jjalcaraz-upct/network-slicing/issues/2, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AN3EW2FE64AKMW335WTCKDLWDCXGPANCNFSM6AAAAAAREZHJJM. You are receiving this because you are subscribed to this thread.Message ID: @.***>