An ADMM Algorithm for Structure Learning in Equilibrium Networks

Department

Program

Citation of Original Publication

Rights

Attribution 4.0 International

Abstract

Learning the edge connectivity structure of networked systems from limited data is a fundamental challenge in many critical infrastructure domains, including power, traffic, and finance. Such systems obey steady-state conservation laws: x(t) = L∗y(t), where x(t) and y(t) ∈ Rᵖ represent injected flows (inputs) and potentials (outputs), respectively. The sparsity pattern of the p x p Laplacian L* encodes the underlying edge structure. In a stochastic setting, the goal is to infer this sparsity pattern from zero-mean i.i.d. samples of y(t). Recent work by Rayas et al. [1] has established statistical consistency results for this learning problem by considering an ℓ₁-regularized maximum likelihood estimator. However, their approach did not focus on developing a scalable algorithm but relies on solving a convex program via the CVX package in Python. To address this gap, we propose an alternating direction method of multipliers (ADMM) algorithm. Our approach is simple, transparent, and computationally fast. A key contribution is demonstrating the role of a non-symmetric algebraic Riccati equation in the primal step of ADMM. Numerical experiments on a host of synthetic and benchmark networks, including power and water systems, show that our method achieves high recovery accuracy.