Accelerated Gradient Methods and Dual Decomposition in Distributed Model Predictive Control

Reference

P. Giselsson, M. D. Doan, T. Keviczky, B. De Schutter, and A. Rantzer, "Accelerated gradient methods and dual decomposition in distributed model predictive control," Automatica, vol. 49, no. 3, pp. 829-833, Mar. 2013.

Abstract

We propose a distributed optimization algorithm for mixed L₁/L₂-norm optimization based on accelerated gradient methods using dual decomposition. The algorithm achieves convergence rate O(1/k²), where k is the iteration number, which significantly improves the convergence rates of existing duality-based distributed optimization algorithms that achieve O(1/k). The performance of the developed algorithm is evaluated on randomly generated optimization problems arising in distributed model predictive control (DMPC). The evaluation shows that, when the problem data is sparse and large-scale, our algorithm can outperform current state-of-the-art optimization software CPLEX and MOSEK.

Publisher page

• DOI link

Downloads

• PDF of corresponding technical report (706 KB)

BibTeX

@article{GisDoa:12-011,
   author  = {Giselsson, Pontus and Doan, Minh Dang and Keviczky, Tam{\'a}s and
              De Schutter, Bart and Rantzer, Anders},
   title   = {Accelerated Gradient Methods and Dual Decomposition in
              Distributed Model Predictive Control},
   journal = {Automatica},
   volume  = {49},
   number  = {3},
   pages   = {829--833},
   month   = mar,
   year    = {2013}
   }

• Download BibTeX entry