Large problems are not necessarily hard: A case study on distributed NMPC paying off

TL;DR

This paper demonstrates that distributed model predictive control can be computationally efficient and scalable for large systems, challenging the assumption that large problems are inherently hard to solve.

Contribution

The study introduces a tailored decentralized real-time iteration scheme for cooperative DMPC, showing it scales well and competes with centralized solvers in power system frequency control.

Findings

DMPC iteration count independent of subsystem number

Decentralized algorithms competitive with centralized solvers

Scalability demonstrated on linear and nonlinear benchmarks

Abstract

A key motivation in the development of Distributed Model Predictive Control (DMPC) is to accelerate centralized Model Predictive Control (MPC) for large-scale systems. DMPC has the prospect of scaling well by parallelizing computations among subsystems. However, communication delays may deteriorate the performance of decentralized optimization, if excessively many iterations are required per control step. Moreover, centralized solvers often exhibit faster asymptotic convergence rates and, by parallelizing costly linear algebra operations, they can also benefit from modern multicore computing architectures. On this canvas, we study the computational performance of cooperative DMPC for linear and nonlinear systems. To this end, we apply a tailored decentralized real-time iteration scheme to frequency control for power systems. DMPC scales well for the considered linear and nonlinear benchmarks, as the iteration number does not depend on the number of subsystems. Comparisons with multi-threaded centralized solvers demonstrate competitive performance of the proposed decentralized optimization algorithms.