Sparsity and Spatial Localization Measures for Spatially Distributed Systems

Abstract

We consider the class of spatially decaying systems, where the underlying dynamics are spatially decaying and the sensing and controls are spatially distributed. This class of systems arise in various applications where there is a notion of spatial distance with respect to which couplings between the subsystems can be quantified using a class of coupling weight functions. We exploit spatial decay property of the underlying dynamics of this class of systems to introduce a class of sparsity and spatial localization measures. We develop a new methodology based on concepts of q-Banach algebras of spatially decaying matrices that enables us to establish a relationship between spatial decay properties of spatially decaying systems and their sparsity and spatial localization features. Moreover, it is shown that the inverse-closedness property of matrix algebras plays a central role in exploiting various structural properties of spatially decaying systems. We characterize conditions for exponentially stability of spatially decaying system over q-Banach algebras and prove that the unique solutions of the Lyapunov and Riccati equations over a proper q-Banach algebra also belong to the same q-Banach algebra. It is shown that the quadratically optimal state feedback controllers for spatially decaying systems are sparse and spatially localized in the sense that they have near-optimal sparse information structures.

Phase retrieval of complex and vector-valued functions

“…associated with a complex conjugate invariant linear space C, the linear space of vector fields on a spatially distributed network [10,18,33,47,48], and vector-valued reproducing kernel spaces in multi-task learning [7,15,42,45]. We say that a family Φ of linear measurements on S is unitary invariant if…”

Network Sparsification with Guaranteed Systemic Performance Measures

“…Let us represent the corresponding coupling graph to L1 by G1. If L1 is obtained via traditional optimal control methods without incorporating sparsity measures, then one should expect to get a dense interconnection topology for G1; we refer to [18] for discussions on spatially decaying structure of optimal controllers. Therefore, our design objective is to compute a localized abstraction for the closed-loop network N(L0 + L1) that only sparsifies G1.…”

“…where c and γ are positive numbers and i, j ∈ V. This class of networks arises in various applications where there is a notion of spatial distance between the subsystems; we refer to [18] for more details. Fig.…”

Differential subalgebras and norm-controlled inversion

“…To consider differential subalgebras of infinite matrices in the noncommutative setting, we introduce three noncommutative Banach algebras of infinite matrices with certain off-diagonal decay. Given 1 ≤ p ≤ ∞ and α ≥ 0, we define the Gröchenig-Schur family of infinite matrices by 25,29,35,43,45], the Baskakov-Gohberg-Sjöstrand family of infinite matrices by 17,22,39,43], and the Beurling family of infinite matrices…”