DRL-based algorithms for solving SFC Embedding Problems
- Apply Advantage Actor-Critic DRL algorithm to solve the SFC embedding problem
- Run "AC_SFC_Main_01.05.py" for training/testing the A2C RL-agent
- Apply DQN RL algorithm to solve the SFC embedding problem
- Run "DQ_SFC_Main_01.08.py" for training the DQN RL-agent
- Run "test_model_random_req_order_v01.08.py" for testing the DQN RL-agent
-
Using the Dataset created by Huy Duong in the Ciena SOF-WP1 Project
-
Notes
- The "reordered_traffic_500000_slots_1_con.tra" file cannot be committed to GitHub due to larger than 100 MB of size. This file will be uploaded to another cloud repo.
- Re-organize each algorithm's source code into a separated folder