rl_channel_allocation

Basic info This is essentially a bin packing problem environment. It is in it's early stages of development but is fully operational.

There are two python scripts in this repistory:

1. environment.py : This is the env which the agent exists in and implements selected actions and calculates the reward earned.
2. qlearn.py : This is the Q-Learning algorithm and handles the acquisition of the calculation of the state-action values.

TODO:

• Enable larger state spaces.
• Allow for dynamic figuring out of actions.
• Find what the best learning rate and discount factor are for this problem.
• Add prioritisation in.
• Add time-awareness.