Machine Learning for Spectrum Access

June 14, 2023

In this project, I implement a machine learning (ML) model for spectrum distribution between multiple users for communication.

The problem I wanna solve using ML is a distributed communication network consisting of multiple users. Distributed communication network means the users are not communicating with each other. The main problem in such a network is that since the users are not connected, their packets might collide and be ruined in the access point (AP). However, users can take advantage of ML to avoid collisions using ML's intelligence in decision making. I take advantage of deep reinforcement learning (DRL) algorightms since the system model is time series.

The algorithm I use is a combination of actor-critic algorithm and double Q-iteration algorithms.

Conclusion

To be updated!