Where the policy iteration of reinforcement learning meets HLL.
Requirements
pip install -r requirements.txt
How it works
- maze.py builds an environment including probability transitions and actions.
- reinforcement_learning.py produces a learning process with a Maze object; no further configuration is needed apart from the user's specification on the grid dimension and learning rate. When the iterations finish, a grid animation including the action mapping from every state (i.e. grid field) is produced as 'animation.gif'.
- qhll.py is an implementation of the quantum version of solving linear systems of equations based on the following paper: https://journals.aps.org/prl/abstract/10.1103/PhysRevLett.103.150502