puzzle15Gym 1.0.1

A custom AI Gym environment for the sliding block puzzle game

Puzzle15Gym

A custom AI Gym environment for the 15-puzzle game: https://en.wikipedia.org/wiki/15_puzzle.

The blank space is represented by -1.

Library used: .

Usage

Initiating the env via gym

import gym
import puzzle15Gym

env_3x3_random = gym.make('Puzzle3x3Random-v0')
env_3x3_fixed = gym.make('Puzzle3x3Fixed-v0')

env_4x4_random = gym.make('Puzzle4x4Random-v0')
env_4x4_fixed = gym.make('Puzzle4x4Fixed-v0')

env_5x5_random = gym.make('Puzzle5x5Random-v0')
env_5x5_fixed = gym.make('Puzzle5x5Fixed-v0')

Initiating the env directly

from puzzle15Gym import Puzzle15Env

env_random = Puzzle15Env(height=4, width=4)
env_random = Puzzle15Env(custom_puzzle="2 8 6|7 1 3|-1 5 4")

Making moves

env_3x3_random = gym.make('Puzzle3x3Random-v0')

# Reset the environment
observation, info = env_3x3.reset()

# Take a step
action = env_3x3.action_space.sample()
observation, reward, done, truncated, info = env_3x3.step(action)

# Render the environment. The only render mode is 'human' which renders visual output.
env_3x3.render()

# Close the environment
env_3x3.close()

Environment Details

• Action Space: Discrete(4) - 0: up, 1: right, 2: down, 3: left.
• Observation Space: Box(-1, height*width-1, (height*width), int32). Contains unique values from -1 to (width * height - 1), excluding 0.
• Reward: 1 if the puzzle is solved, 0 if not, -2 if invalid move.
• Done: True if the puzzle is solved, False otherwise.