Mathematics of Deep Learning
Spring 2020

Iterative linear quadratic regulation


Iterative linear quadratic regulation (iLQR) approximates the dynamics using a time-varying linear model and approximately solves it using an iterative algorithm. It enables optimal control via trajectory optimization for arbitrary environments where the dynamics are known or can be approximated. Guided Policy Search uses iLQR to find optimal guiding trajectories. After this week you should understand how iLQR solves nonlinear trajectory optimization problems.


Required reading

Optional reading


References from the session and further resources