search

baturaysaglam / RIS-MISO-Deep-Reinforcement-Learning

Joint Transmit Beamforming and Phase Shifts Design with Deep Reinforcement Learning
MIT License
130 stars 37 forks source link

Power conditions used to create figures 8, 11 and 12 #16

Open Mubasher-KHU opened 10 months ago

Mubasher-KHU commented 10 months ago

Hi, I was trying to reproduce the results in the paper but I am unable to get the same results. Would it be possible for your to share the power that was used to create the plots for figures 8, 11 and 12 as the final reward in my case is not matching with the figures. Thank you

baturaysaglam commented 10 months ago

sorry, it's been a while since I reproduced these results. I can't remember the tested configurations exactly to be honest. did you check the original figures reported in the paper corresponding to these figures? did the authors mention any power level?

Mubasher-KHU commented 10 months ago

Ahh okay, thanks. I checked the paper but the authors have not mentioned the power level. Also, I am running the code but it is taking much for steps for me to reach an acceptable level of reward. I had to run the algorithm for 50000 steps per episode. I did not change any of the hyperparameters and the channel setting was also taken to be the same as the paper. Any tips on what I can check to get similar results in fewer steps? Thank you

Always1172 commented 9 months ago

Hi ,I meet the same issue as well, I tried 1000,2000,5000,10000,20000steps per episode, but can't get the ideal reward. Did you find solutions? Thank you

MuhammadAbulhassan900 commented 6 months ago

Hi, I am bit confused about the .npy data in each learning Curves/power etc, how you extract such .npy files. I execute as it is but still unable to generate such files to generate figures. Any one how is actively following this repository is welcome to guide me. Thanks

Deemah42 commented 5 months ago

@MuhammadAbulhassan900 I have the same question. I executed the results for 350 episodes, 10000 steps, but I still do not get how the .npy files are one vector of 10000 steps.