Feedback Controller Sparsification for a Class of Linear Systems with Parametric Uncertainties

TL;DR

This paper presents a method for sparsifying output feedback controllers in linear systems with parametric uncertainties, balancing performance preservation and sparsity, and demonstrating effectiveness on a power network example.

Contribution

It introduces an optimization framework that minimizes performance loss during sparsification, reformulates the problem into a rank-constrained optimization, and provides an iterative algorithm for sub-optimal solutions.

Findings

Effective sparsification with minimal performance degradation

Successful application to IEEE 39-bus power network

Reformulation into rank-constrained optimization enhances solution quality

Abstract

We consider the problem of output feedback controller sparsification for systems with parametric uncertainties. We develop an optimization scheme that minimizes the performance deterioration caused by the sparsification process, while enhancing sparsity pattern of the feedback gain. In order to improve temporal proximity of an existing closed-loop system and its sparsified counterpart, we also incorporate an additional constraint into the problem formulation so as to bound the variation in the system output pre and post sparsification. We also show that the resulting non-convex optimization problem can be equivalently reformulated into a rank-constrained optimization problem. We then formulate a bi-linear minimization program along with an iterative algorithm to obtain a sub-optimal solution which satisfies the rank constraint with arbitrary tolerance. Lastly, a sub-optimal sparse controller design for IEEE 39-bus New England power network is utilized to showcase the effectiveness of our proposed method.