December 2019
Intermediate to advanced
368 pages
11h 10m
English
The simulation environment must provide minimal information about the system state after performing actions in the form of a reinforcement signal (
). The reinforcement signal indicates whether a double-pole balancing system violates the bounding constraints after applying an action. It can be defined as follows:

The implementation of the reinforcement signal generation in Python is as follows:
res = x < -2.4 or x > 2.4 or \ theta1 < -THIRTY_SIX_DEG_IN_RAD or theta1 > THIRTY_SIX_DEG_IN_RAD or \ theta2 < -THIRTY_SIX_DEG_IN_RAD ...
Read now
Unlock full access