**Reference:**

P. Giselsson,
M.D. Doan,
T. Keviczky,
B. De Schutter, and
A. Rantzer,
"A distributed optimization algorithm with convergence rate
O(1/k^{2}) for distributed model predictive control," Tech.
rep. 12-011, Delft Center for Systems and Control, Delft University of
Technology, Delft, The Netherlands, Mar. 2012.

**Abstract:**

We propose a distributed optimization algorithm for mixed
L_{1}/L_{2}-norm optimization based on accelerated
gradient methods using dual decomposition. The algorithm achieves
convergence rate O(1/k^{2}), 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 (MPC). The evaluation shows that, when the problem
data is sparse and large-scale, our algorithm outperforms
state-of-the-art optimization software CPLEX and MOSEK.

Technical report: pdf file (171 KB)

@techreport{GisDoa:12-011,

author={P. Giselsson and M.D. Doan and T. Keviczky and B. {D}e Schutter and A. Rantzer},

title={A distributed optimization algorithm with convergence rate ${O}(\frac{1}{k^2})$ for distributed model predictive control},

number={12-011},

institution={Delft Center for Systems and Control, Delft University of Technology},

address={Delft, The Netherlands},

month=mar,

year={2012}

}

Go to the publications overview page.

This page is maintained by Bart De Schutter. Last update: December 15, 2015.