Distributed optimal control of nonlinear systems using a second-order augmented Lagrangian method

In this paper, we propose a distributed second-order augmented Lagrangian method for distributed op-timal control problems, which can be exploited for distributed model predictive control. We employ a primal-dual interior-point approach for the inner iteration of the augmented Lagrangian and distribute the corresponding computations using message passing over what is known as the clique tree of the problem. The algorithm converges to its centralized counterpart and it requires fewer communications between sub-systems as compared to algorithms such as the alternating direction method of multipli-ers. We illustrate the efficiency of the framework when applied to randomly generated interconnected sub-systems as well as to a vehicle platooning problem.(c) 2023 The Author(s). Published by Elsevier Ltd on behalf of European Control Association. This is an open access article under the CC BY license ( http://creativecommons.org/licenses/by/4.0/ )

Automated summary

In this paper, a distributed second-order augmented Lagrangian method is proposed for distributed optimal control problems, particularly for distributed model predictive control. A primal-dual interior-point approach is used for the inner iteration of the augmented Lagrangian, and computations are distributed using message passing over the clique tree of the problem. The algorithm converges to its centralized counterpart and requires fewer communications between sub-systems compared to other algorithms like the alternating direction method of multipliers. The efficiency of the framework is demonstrated through applications on randomly generated interconnected sub-systems and a vehicle platooning problem.