def getRandom3DInstance(self, with_velocity=True):
env = ThreeDMountainCarEnv()
env.reset()
random_pos_x = np.random.uniform(low=env.min_position_x,
high=env.max_position_y)
random_pos_y = np.random.uniform(low=env.min_position_x,
high=env.max_position_y)
# TODO: calculates the maximum speed at this position
random_velocity_x = np.random.uniform(
low=-env.max_speed_x, high=env.max_speed_x) if with_velocity else 0
random_velocity_y = np.random.uniform(
low=-env.max_speed_y, high=env.max_speed_y) if with_velocity else 0
state = [
random_pos_x, random_pos_y, random_velocity_x, random_velocity_y
]
env.set_state(state)
action = np.random.randint(low=0, high=5)
next_state, reward, done, info = env.step(action)
return [state, action, next_state, reward, done]
def main():
# env = gym.make("MountainCar-v0")
env = ThreeDMountainCarEnv()
act = deepq.load("mountaincar_model_working.pkl")
while True:
obs, done = env.reset(), False
episode_rew = 0
while not done:
# env.render()
env.render_orthographic()
obs, rew, done, _ = env.step(act(obs[None])[0])
print(act(obs[None])[0])
print(obs)
episode_rew += rew
print("Episode reward", episode_rew)
import gym
import numpy as np
from matplotlib import pyplot as plt
import itertools
from lib.env.threedmountain_car import ThreeDMountainCarEnv
env = ThreeDMountainCarEnv()
state = env.reset()
for t in itertools.count():
# action = env.action_space.sample()
next_state, reward, done, info = env.step(0)
# env.render() # yellow
# env.render_y() #cyan
env.render_orthographic()
if done:
break
state = next_state
# if t == 100:
# env.close_gui()
# break