Value of Communication: Data-Driven Topology Optimization for Distributed Linear Cyber-Physical Systems

TL;DR

This paper introduces a data-driven approach to optimize communication topology in distributed cyber-physical systems, enhancing control performance by selectively designing network links without requiring a detailed system model.

Contribution

It presents a novel mixed-integer second-order conic programming method for topology design that improves control outcomes and reduces unnecessary communication links.

Findings

Optimized topologies outperform random configurations in control tasks.

The method effectively identifies and removes links with minimal impact on control accuracy.

Control cost correlates strongly with predictor accuracy, guiding link removal.

Abstract

Communication topology is a crucial part of a distributed control implementation for cyber-physical systems, yet is typically treated as a constraint within control design problems rather than a design variable. We propose a data-driven method for designing an optimal topology for the purpose of distributed control when a system model is unavailable or unaffordable, via a mixed-integer second-order conic program. The approach demonstrates improved control performance over random topologies in simulations and efficiently drops links which have a small effect on predictor accuracy, which we show correlates well with closed-loop control cost.