Gridworld

KO | EN

gitlite — search

#javascript #python #ai #typescript #rust #linux #macos #nodejs #java #awesome #cli #kubernetes #awesome-list #open-source

(★ 15)

Simple implementation of text-based Gridworld game. Intended for use with reinforcement learning algorithms.

Gridworld 최신버젼 다운로드

최종 버전 다운로드 (.zip)

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# Gridworld
Simple implementation of text-based Gridworld game. Intended for use with reinforcement learning algorithms.

As described in this blog post: < http://outlace.com/rlpart3.html >

### Basic usage:

```python
from Gridworld import *
```

Initialize game state using `initGrid` (initializes stationary grid, all items are placed deterministically), `initGridPlayer` (initializes player in random location, but keep wall, goal and pit stationary), or `initGridRand` in which all pieces are initialized to random locations on the board.

```python
game = Gridworld(size=4, mode='static')
game.dispGrid()
```

```
array([['+', '-', ' ', 'P'],
       [' ', 'W', ' ', ' '],
       [' ', ' ', ' ', ' '],
       [' ', ' ', ' ', ' ']], dtype='<U2')

```

"P" represents the player, "W" is the wall (an obstacle), and "-" is the pit (a trap that gives a negative reward) and "+" is the goal which gives positive reward and wins the game.

The player makes moves with the `game.makeMove(move)` where  `move` is a character `u,d,l, or r` for up,down, left or right.

```python
game.makeMove('d') # (0,3) + (-1,3)
print("Reward: %s" % (game.getReward()))
game.dispGrid()
```

```
Reward: -1

array([['+', '-', ' ', ' '],
       [' ', 'W', ' ', 'P'],
       [' ', ' ', ' ', ' '],
       [' ', ' ', ' ', ' ']], dtype='<U2')
```

Use the `render_np()` method to produce a numpy tensor to supply to a machine learning algorithm.
```python
game.board.render_np()
```

```
array([[[0, 0, 0, 0],
        [0, 0, 0, 1],
        [0, 0, 0, 0],
        [0, 0, 0, 0]],

[[1, 0, 0, 0],
        [0, 0, 0, 0],
        [0, 0, 0, 0],
        [0, 0, 0, 0]],

[[0, 1, 0, 0],
        [0, 0, 0, 0],
        [0, 0, 0, 0],
        [0, 0, 0, 0]],

[[0, 0, 0, 0],
        [0, 1, 0, 0],
        [0, 0, 0, 0],
        [0, 0, 0, 0]]], dtype=uint8)

```

Gridworld

Gridworld 최신버젼 다운로드

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

Gridworld

Gridworld 최신버젼 다운로드

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

Download README