October 10, 2025Open Access

A Sequential Quadratic Programming Perspective on Optimal Control

View Full Paper

Ask AI

Bookmark

Authors

Discussion

This analysis demonstrates the connection between iLQR and SQP methods, suggesting improved convergence for optimal control tasks.

The iLQR method is a principled application of sequential quadratic programming, ensuring a cost-descent direction.
Empirical evaluations on pendulum and cart-pole tasks validate the Sequential Quadratic Programming approach.
Newton's method and DDP lack convergence guarantees far from an optimum, emphasizing the advantage of iLQR.
This unified perspective on optimal control highlights the relationships between various iterative approaches.

Abhijeet et al. (2025) studied this question.

View Full Paper

Ask AI

Bookmark