This repository contains a PIP package which is an OpenAI environment for simulating an wumpusworld and fuzzy-cartpole environment.

Install the OpenAI gym.

Then install this package via

pip install -e .

or

pip install git+https://github.com/spelfrog/ai-gym.git@master#egg=fh-ac-ai-gym

import gym
import fh_ac_ai_gym

wumpus_env = gym.make('Wumpus-v0')
cartpole_env = gym.make('Fuzzy-CartPole-v0')

The wumpus action space is containing the following actions:

   WALK = 0
   TURNLEFT = 1
   TURNRIGHT = 2
   GRAB = 3
   SHOOT = 4
   CLIMB = 5

The fuzzy env is taking the applied force +/- to the cart as action. For the cartpole env you have to use .step_with_(force) instead of the step(action)

   obs, reward,done, info = env.step_with_(cart.output['force'])