Rank minimization and applications in system theory

Abstract

In this tutorial paper, we consider the problem of minimizing the rank of a matrix over a convex set. The rank minimization problem (RMP) arises in diverse areas such as control, system identification, statistics and signal processing, and is known to be computationally NP-hard. We give an overview of the problem, its interpretations, applications, and solution methods. In particular, we focus on how convex optimization can be used to develop heuristic methods for this problem.