Hands-On Neuroevolution with Python

The system state and equations of motion

The goal of the controller is to apply the force to the cart in order to keep two poles balanced for as long as possible. At the same time, the cart should stay within the defined boundaries. As with the single-pole balancing problem discussed previously, the control strategy can be defined as an avoidance control problem, which means that the controller must maintain a stable system state, avoiding the danger zones when the cart moves outside the track boundaries or either of the poles fall beyond the allowed vertical angle. There is no unique solution for this problem, but an appropriate control strategy can be found because the poles have different lengths and masses. Therefore, they respond differently ...