/DQN_MazeSolver

Find a path to the end in the maze using Noisy Deep Q networks.

Primary LanguagePythonMIT LicenseMIT

DQN Maze

Release Tag

Issues Commits

Languages Size

The topology of the NoisyNet.

model

In hidden layers is used ReLU activation function and linear activation function is used in the output layer.

States

26 inputs = 2 position + 24 objects around agent

Actions

  • Up
  • Down
  • Left
  • Right

Run

python3 main.py

License

MIT