Tabular Q, DQN and DDQN for movement of a cursor (represents the hand) between two points as fast as possible.
Reaching Task
- The task is to move the cursor from start state (position, velocity) = (0,0) to a target state (position, velocity) = (8,0).
- The agent chooses between one of the three acceleration actions (-1, 0, 1) at each step of movement.