@@ -67,6 +67,17 @@ To make sure that you have the necessary libraries, you can create a virtual env
- my_example: an installation folder for RL-VNE implementation using openAI gym interface.
- ToyExample.ipynb: an example on how to use the gym environment.
It is basically enough to use RL_VNE_Notebook.ipynb to understand the paper's implementation
It is basically enough to use [RL_VNE_Notebook.ipynb](../master/RL_VNE_Notebook.ipynb) to understand the paper's implementation
## General Comments
RL provide fast solutions with very good results (of course they depend on the problem formulation). This is an introductory work for using RL in solving VNE problems. The aim is to provide simple and easy-to-use implementation that can be solved with different algorithms (e.g., algorithms from `stable baselines` framework), or even be extended to solve more complicated VNE problems.
In an extended version of this paper, we plan to use other algorithms and tune their hyperparemeters for faster convergence and probably lower gap to the optimization model.
To build up on this work, we will explore similar problems but with more complicated scenarios (e.g., dynamic nodes instead of power-plugged ones).
Extention is in progress. Links will be provided once primary results are available.
