Input Matrix Construction and Approximation Using a Graphic Approach

TL;DR

This paper introduces a graph-theoretic approach to construct and approximate the sparsest input matrices for controllability in state transition matrices, providing polynomial algorithms and approximation guarantees.

Contribution

It offers a new characterization for input matrix sparsity patterns, extends controllability results, and develops efficient algorithms with provable guarantees for sparse control design.

Findings

Polynomial-time procedure for constructing controllable input matrices.

Minimal number of inputs equals maximum geometric multiplicity under certain constraints.

Two-stage algorithm with provable approximation guarantees for sparsity minimization.

Abstract

Given a state transition matrix (STM), we reinvestigate the problem of constructing the sparest input matrix with a fixed number of inputs to guarantee controllability. We give a new and simple graph theoretic characterization for the sparsity pattern of input matrices to guarantee controllability for a general STM admitting multiple eigenvalues, and provide a deterministic procedure with polynomial time complexity to construct real valued input matrices with arbi- trarily prescribed sparsity pattern satisfying controllability. Based on this criterion, some novel results on sparsely controlling a system are obtained. It is proven that the minimal number of inputs to guarantee controllability equals to the maximum geometric multiplicity of the STM under the constraint that some states are actuated-forbidden, extending the results of [28]. The minimal sparsity of input matrices with a fixed number of inputs is not necessarily equal to the minimal number of actuated states to ensure controllability. Furthermore, a graphic sub- modular function is built, leading to a greedy algorithm to efficiently approximate the minimal actuated states to assure controllability for general STMs. For the problem of approximating the sparsest input matrices with a fixed number of inputs, we propose a simple greedy algo- rithm (non-submodular) and a two-stage algorithm, and demonstrate that the latter algorithm, inspired from techniques in dynamic coloring, has a provable approximation guarantee. Finally, we present numerical results to show the efficiency and effectiveness of our approaches.