Skip to content

Instantly share code, notes, and snippets.

@jogam5

jogam5/wwmse_v1.txt

Last active December 8, 2021 05:36
Show Gist options
  • Save jogam5/d1a91163ea54bee497bf1d618a2acf58 to your computer and use it in GitHub Desktop.
Save jogam5/d1a91163ea54bee497bf1d618a2acf58 to your computer and use it in GitHub Desktop.
Download ZIP
wwmse example
Raw
wwmse_v1.txt
# Note: 1 and 2 are simplifications to be more clear
# 1. Generate distribution: simplification
data_from_distribution = generate_rayleigh_CSI(parameter_1, parameter_2, parameter_3, ...)
# 2 Loop over data to compute wwmse: simplification
for sample in range(data_from_distribution):
wwmse = WMMSE(number_of_transceivers, sample, max_power_budget, noise)
# ==Parameters==:
#
# number_of_transceivers: number of single-antenna
# transceivers pairs (type: ndarray, shape: (10,), default=10)
# sample: sample of the distribution to be computed (type: ndarray, shape: (10,10))
# max_power_budget: power budget of each transmitter (type: int, default:1)
# noise: noise (type: float, default:1.0)
# =========================
# Note: 3 and 4 are real code
# 3. Output from one of the different distributions methods (rayleigh) before
# entering the loop. When it enters the loop, the array (1,100) is reshaped into a
# (10,10) array. The (10,10) array is the one used as input for parameter "H".
#
Rayleigh: (20000, 100) , <class 'numpy.ndarray'>
Rayleigh first row exammple (1, 100): [0.73726645 0.4591307 1.17678047 1.06890929 1.7022319 0.70239894
1.00141185 1.28763866 1.07092557 0.08352788 0.76777986 1.38532127
0.42888726 0.27903338 0.58691025 0.69247862 1.38472584 1.31138042
0.33567388 0.66638627 0.95206096 1.05981347 2.24319029 0.84454063
1.04057974 0.40758808 0.9311077 0.56890818 0.55937325 0.89496559
0.8980085 0.83251139 0.58396146 0.97541061 1.26303782 1.32623366
0.54001067 0.58822374 0.41641521 0.31058773 0.50976306 0.39001463
1.70138358 1.74412918 0.95101006 0.36662434 0.2295487 0.59322144
0.54342352 0.33651321 2.34283749 1.21255351 0.51790161 0.70098649
0.60306511 0.87564781 1.50796222 0.26442621 1.04125186 0.36776211
0.71884519 0.92494189 0.58118326 1.27835902 0.22412188 1.45130246
0.86711031 0.35353472 1.28052904 0.82492686 0.33711347 0.93950375
0.84693652 0.9374323 1.99400671 1.71494858 0.69775086 0.16599246
0.79118827 0.8957736 0.39370973 1.20624926 1.61031853 0.27995009
0.98711437 1.67149709 1.13517947 0.62234045 0.85480029 0.57429993
0.39867545 1.65346519 0.54472579 1.69342232 0.26588566 0.92129592
0.17996725 1.22113943 0.8252 0.90242043]
# 4. Inside wmmse
def WMMSE_sum_rate(p_int, H, Pmax, var_noise):
"""
Returns the Weighted System Throughput based on the
WMMSE algorithm [3].
The WMMSE algorithm converts the weighted sum-rate (WSR)
maximization problem to a higher dimensional space where
it can be solved. Following [2], the algorithm was modified
to work in the real domain.
Input
# variables/values can be found in channel.py and generate_data.py file
p_int: number of single-antenna transceivers pairs (type: ndarray, shape: (10,), default: 10)
H: sample of the distribution to be computed (type: ndarray, shape: (10,10))
Pmax: power budget of each transmitter (type: int, default:1)
var_noise: noise (type: float, default:1.0)
Output
p_opt: weighted system throughput (type: ndarray, shape: (10,10))
"""
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment