Implementation of a simple game along with an AI agent, able to learn to play the game.
The AI agent is an implementation of a Deep-Q-Network using TensorFlow.
In this game, our character has to move in a 2D grid in order to find the goal while avoiding to fall in the hole.
In this grid, the character can either move up/down/right/left. Moving toward the edge of the grid moves the character on the other side of the grid.
exemple of grid:
X - -
- - -
- O -
- - 1
X: character
0: hole
1: goal