An efficient virtualized network function deployment scheme for service function chain using deep Q-network
File Descriptions
DQN-QQE
-
main.py
- It is the main function that executes DQN-QQE.
- Moreover, random and brute force are also executed in main.py.
-
env.py
- This file sets the environment of DQN-QQE.
-
dqn.py
- There are the neural network structure and lots of DQN functions.
-
energy.py
- It contains the energy process.
-
config.py
- It sets lots of parameters.
- DQN-QQE and DQN-Q2-SFC use the same file.
DQN-Q2-SFC
-
main_QQ.py
- It is the main function that executes DQN-QQE.
- Moreover, random and brute force are also executed in main.py.
-
env_QQ.py
- This file sets the environment of DQN-QQE.
-
dqn_QQ.py
- There are the neural network structure and lots of DQN functions.
-
config.py
- It sets lots of parameters.
- DQN-QQE and DQN-Q2-SFC use the same file.
Random
- random_sfc.py
- Random scheme.
Brute Force
-
brute_sfc_energy.py
- Brute force approach on energy aspect.
-
brute_sfc_qoe.py
- Brute force approach on qoe aspect.
Execute
Run run.bat
- It is a batch file.
- Line 1:
call C:\Users\SNMLAB\Anaconda3\Scripts\activate.bat C:\Users\SNMLAB\Anaconda3- It need to know where the activate.bat of Anaconda3 is.
- Line 2:
python main.py- Run DQN-QQE, random, and brute force.
- Line 3:
python main_QQ.py- Run DQN-Q2-SFC
- Line 4:
pause- Stop batch file.
Plots
Quality of Experience (QoE)
-
plot_qoe.py
- Line chart
-
plot_qoe_boxplot.py
- Boxplot chart
Error rate
-
plot_error.py
- Line chart
-
plot_error_boxplot.py
- Boxplot chart
Maximum energy consumption
-
plot_max_energy.py
- Line chart
-
plot_max_energy_boxplot.py
- Boxplot chart
Average energy consumption
-
plot_mean_energy.py
- Line chart
-
plot_mean_energy_boxplot.py
- Boxplot chart
Average processing time
-
plot_time.py
- Line chart
-
plot_time_barplot.py
- Bar chart
Using Ant Colony Optimization algorithm (ACO) to iris dataset
- To get the clusters
- ACO.ipynb