Core Concepts
The authors propose a novel control-theoretic approach to design online algorithms that can efficiently solve constrained optimization problems with time-varying costs and constraints, achieving zero tracking error.
Abstract
The authors focus on solving online optimization problems with time-varying, linear equality and inequality constraints. They take a control-theoretic approach to design novel online algorithms that can track the optimal trajectory with zero error, in contrast to alternative unstructured and structured methods that only achieve non-zero tracking error.
For problems with only equality constraints, the authors design an algorithm that leverages the internal model principle and robust control techniques to achieve asymptotic convergence to the optimal trajectory. When inequality constraints are also present, the authors extend the algorithm by incorporating an anti-windup mechanism to handle the non-negativity constraints on the dual variables.
The authors provide numerical results that demonstrate the superior performance of the proposed algorithms compared to state-of-the-art methods, both for quadratic and non-quadratic cost functions, and in the presence of equality and inequality constraints.
Stats
The authors report numerical results that corroborate the theoretical analysis and show how the proposed approach outperforms state-of-the-art algorithms both with equality and inequality constraints.