Package kuimaze :: Module gym_wrapper
[hide private]
[frames] | no frames]

Module gym_wrapper

source code

File wrapping maze.py functionality into few methods defined by OpenAI GYM


Author: Zdeněk Rozsypálek, Tomas Svoboda

Contact: rozsyzde(at)fel.cvut.cz, svobodat@fel.cvut.cz

Copyright: (c) 2017, 2018

Classes [hide private]
  MazeEnv
  EnvAgent
Class necessary for wrapping maze
  EasyMazeEnv
EasyMazeEnv is version of maze closer to graph search.
  InfEasyMaze
informed easy maze, suitable for A* implementation step([x, y])
  EasyMaze
uninformed easy maze, suitable for BFS, DFS ...
  MDPMaze
maze for solving MDP problems
  HardMaze
Uninformed hard maze, suitable for reinforcement learning step(param) where param is integer; 0 <= param <= 3
  InfHardMaze
Informed hard maze, suitable for reinforcement learning step(param) where param is integer; 0 <= param <= 3
Variables [hide private]
  path_section = collections.namedtuple('Path', ['state_from', '...
  state = collections.namedtuple('State', ['x', 'y'])
Variables Details [hide private]

path_section

Value:
collections.namedtuple('Path', ['state_from', 'state_to', 'cost', 'act\
ion'])