Accelerated gradient methods and dual decomposition in distributed model predictive control

Authors:
Pontus Giselsson;Minh Dang Doan;TamáS Keviczky;Bart De Schutter;Anders Rantzer
Affiliations:
Department of Automatic Control, Lund University, Sweden;Delft Center for Systems and Control, Delft University of Technology, The Netherlands;Delft Center for Systems and Control, Delft University of Technology, The Netherlands;Delft Center for Systems and Control, Delft University of Technology, The Netherlands;Department of Automatic Control, Lund University, Sweden
Venue:
Automatica (Journal of IFAC)
Year:
2013

Citing 5
Cited 0

Applied mathematics in water supply network management

Automatica (Journal of IFAC) - IFAC-IEEE special issue on meeting the challenge of computer science in the industrial applications of control
Smooth minimization of non-smooth functions

Mathematical Programming: Series A and B
Multi-agent model predictive control for transportation networks: Serial versus parallel schemes

Engineering Applications of Artificial Intelligence
A Fast Iterative Shrinkage-Thresholding Algorithm for Linear Inverse Problems

SIAM Journal on Imaging Sciences
A Fast Iterative Shrinkage-Thresholding Algorithm for Linear Inverse Problems

SIAM Journal on Imaging Sciences

Quantified Score

Hi-index	22.14

Visualization

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(1k^2), where k is the iteration number, which significantly improves the convergence rates of existing duality-based distributed optimization algorithms that achieve O(1k). 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.