Can anyone recommend a library or framework of reinforcements that can handle large state spaces, abstracting them?
I am trying to implement intelligence for a small agent in the game world. The agent is represented by a small two-wheeled robot that can move back and forth and rotate left and right. It has a pair of sensors for detecting boundaries on the ground, a pair of ultrasonic sensors for detecting objects far away, and a pair of sensors for detecting contact with an object or an adversary. He can also make some simple calculations to evaluate his position in the world, using his starting position as a reference. Thus, all state functions available to him:
edge_detected=0|1
edge_left=0|1
edge_right=0|1
edge_both=0|1
sonar_detected=0|1
sonar_left=0|1
sonar_left_dist=near|far|very_far
sonar_right=0|1
sonar_right_dist=near|far|very_far
sonar_both=0|1
contact_detected=0|1
contact_left=0|1
contact_right=0|1
contact_both=0|1
estimated_distance_from_edge_in_front=near|far|very_far
estimated_distance_from_edge_in_back=near|far|very_far
estimated_distance_from_edge_to_left=near|far|very_far
estimated_distance_from_edge_to_right=near|far|very_far
, , , , . , , 2985984 , , - Q- SARSA.
- , ( Python) , ?