Exploiting Block Structures of KKT Matrices for Efficient Solution of Convex Optimization Problems

Convex optimization solvers are widely used in the embedded systems that require sophisticated algorithms including model predictive control (MPC). In this paper, we aim to reduce online solve time of such convex so as total runtime algorithm and make it suitable for real-time optimization. We exploit property Karush–Kuhn–Tucker (KKT) matrix involved solution problem only some parts change during iterations algorithm. Our results show proposed method can effectively solvers.