Physics-informed structured learning of a class of recurrent neural networks with guaranteed properties

TL;DR

This paper introduces a physics-informed learning framework for recurrent neural networks that guarantees structural and stability properties, enabling scalable and modular control-oriented modeling of large-scale systems.

Contribution

It develops a convex optimization-based learning method incorporating linear matrix inequalities to enforce system properties, with parallelization for modular systems.

Findings

Effective preservation of stability and structure in learned models

Scalable parallel optimization for large systems

Successful simulation validation of the approach

Abstract

This paper proposes a physics-informed learning framework for a class of recurrent neural networks tailored to large-scale and networked systems. The approach aims to learn control-oriented models that preserve the structural and stability properties of the plant. The learning algorithm is formulated as a convex optimisation problem, allowing the inclusion of linear matrix inequality constraints to enforce desired system features. Furthermore, when the plant exhibits structural modularity, the resulting optimisation problem can be parallelised, requiring communication only among neighbouring subsystems. Simulation results show the effectiveness of the proposed approach.