GymRL

GymRL is a package that provides various environments for reinforcement learning.

Prerequisites

Before you begin, ensure you have met the following requirements:

• You have installed Node.js

Installation

Install GymRL using npm:

$ npm install gymrl

Environments

GymRL currently supports the following environments:

1. TwoThousandfortyeight
2. Fourinarow
3. Frozenlake
4. Snake
5. TicTacToe
6. Rubik
7. Pong

You can import an environment as follows:

CommonJS

const { Snake } = require("gymrl");

ESM

import { TwoThousandfortyeight } from "gymrl";

Usage

Here is an example of how to use the Snake environment:

import { Snake } from "gymrl";

const match = new Snake(10); // This creates a map of 10x10
let score = 0;
let done = false;

match.render();

while (!done) {
  const randomAction = match.sampleAction();
  const Stepinfo = match.step(randomAction);
  const reward = Stepinfo[1];
  done = Stepinfo[2];
  score += reward;
  match.render();
  console.log(`Reward: ${reward}, Total Score: ${score}`);
}

Methods

The following methods are available in all environments:

• render(): Creates a string representation of the environment and displays it in the console.
• sampleAction(): Returns a random action from the environment.
• get_obs(): Returns the current observations of the environment.
• get_info(): Returns additional information about the observations of the environment.
• reset(): Resets the environment.
• step(action): Performs one step in the environment using the specified action and returns an array with four entries: [newStatus, Reward, Done, Info].

More Info About the Environments

You can see more info about the enviroments in our wiki

Contributing to GymRL

If you want to contribute to GymRL, whether by adding a new environment or fixing bugs, please check out the source code and submit a pull request or issue.