A Distributed Active Set Method for Model Predictive Control

TL;DR

This paper introduces a distributed active set method for model predictive control that ensures primal feasibility and competes with existing methods in communication efficiency, especially for linear systems.

Contribution

It combines a primal active set strategy with a decentralized conjugate gradient method, providing a novel approach with primal feasibility in distributed MPC.

Findings

Method maintains primal feasibility of iterates.

Competitively reduces communication in distributed MPC.

Performs well in chain of masses example.

Abstract

This paper presents a novel distributed active set method for model predictive control of linear systems. The method combines a primal active set strategy with a decentralized conjugate gradient method to solve convex quadratic programs. An advantage of the proposed method compared to existing distributed model predictive algorithms is the primal feasibility of the iterates. Numerical results show that the proposed method can compete with the alternating direction method of multipliers in terms of communication requirements for a chain of masses example.