Time Limit Wrapper

Enviroment wrapper to set a maximum number of steps per episode. If the agent does not succesfully end the task or reach the goal the episode will end and the property is_success will be set to 0.

Example use

from kobuki_maze_rl.task_env import kobuki_maze
from frobs_rl.common import ros_gazebo
import gym
import rospy

from frobs_rl.wrappers.TimeLimitWrapper import TimeLimitWrapper

if __name__ == '__main__':

    # Launch Gazebo
    ros_gazebo.launch_Gazebo(paused=True, gui=False)

    # Start node
    rospy.logwarn("Start")
    rospy.init_node('kobuki_maze_train')

    # Launch the task environment
    env = gym.make('KobukiMazeEnv-v0')

    #--- Set max steps
    env = TimeLimitWrapper(env, max_steps=15000)

Wrapper documentation

class TimeLimitWrapper.TimeLimitWrapper(env, max_steps=100)

Wrapper to limit the number of steps per episode.

Parameters

• env – (gym.Env) Gym environment that will be wrapped

• max_steps – (int) Max number of steps per episode

reset()

Reset the environment

step(action)

Parameters

action ([float] or int) – Action taken by the agent

Returns

observation, reward, is the episode over, additional informations

Return type

(np.ndarray, float, bool, dict)