Deep Neural Network for Computation Rate Maximization in Wireless-powered Mobile Edge Computing
Python code to reproduce our works on Wireless-powered Mobile Edge Computing [1], which uses the wireless channel gains as the input and the binary computing mode selection results as the output of a deep neural network (DNN). It includes:
-
dnn_wpmec.py: the DNN structure for the WPMEC, inclduing training structure and test structure
-
data: all data are stored in this subdirectory, includes:
-
data_#.mat: training and testing data sets, where # is the user number
-
Prediction_#.mat: the predicted mode selection generated by DNN_test
-
weights_biases.mat: parameters of the trained DNN, which are used to re-produce this trained DNN such as in MATLAB.
-
-
main.py: run this file, inclduing setting system parameters
Please refer to our recent advantages in this topic as released at DROO. Specifically, a Reinforcement learning-based online algorithm, DROO, is proposed to maximize the weighted compuation rate in Wireless Powered Mobile-Edge Computing Networks. Numerical results show that the proposed algorithm can achieve near-optimal performance while significantly decreasing the computation time by more than an order of magnitude compared with existing optimization methods. For example, the CPU execution latency of DROO is less than 0.1 second in a 30-user network, making real-time and optimal offloading truly viable even in a fast fading environment.
- Liang Huang, Suzhi Bi, and Ying-jun Angela Zhang, "Deep Reinforcement Learning for Online Computation Offloading in Wireless Powered Mobile-Edge Computing Networks", on arxiv:1808.01977.
-
Suzhi BI, bsz AT szu.edu.cn
-
Liang HUANG, lianghuang AT zjut.edu.cn
-
Ying Jun (Angela) Zhang, yjzhang AT ie.cuhk.edu.hk
-
Tensorflow
-
numpy
-
scipy
A brief manual on how to install tensorflow and related packages is provided in English and 简体中文. Contact Liang Huang (lianghuang AT zjut.edu.cn) if you need further help.
run the file, main.py