Conference paper
Input-constrained model predictive control via the alternating direction method of multipliers
Department of Applied Mathematics and Computer Science, Technical University of Denmark1
Scientific Computing, Department of Applied Mathematics and Computer Science, Technical University of Denmark2
Center for Energy Resources Engineering, Centers, Technical University of Denmark3
CITIES - Centre for IT-Intelligent Energy Systems, Centers, Technical University of Denmark4
This paper presents an algorithm, based on the alternating direction method of multipliers, for the convex optimal control problem arising in input-constrained model predictive control. We develop an efficient implementation of the algorithm for the extended linear quadratic control problem (LQCP) with input and input-rate limits.
The algorithm alternates between solving an extended LQCP and a highly structured quadratic program. These quadratic programs are solved using a Riccati iteration procedure, and a structure-exploiting interior-point method, respectively. The computational cost per iteration is quadratic in the dimensions of the controlled system, and linear in the length of the prediction horizon.
Simulations show that the approach proposed in this paper is more than an order of magnitude faster than several state-of-the-art quadratic programming algorithms, and that the difference in computation time grows with the problem size. We improve the method further using a warm-start procedure.
| Language: | English |
|---|---|
| Publisher: | IEEE |
| Year: | 2014 |
| Pages: | 115-120 |
| Proceedings: | 13th European Control Conference (ECC) 2014European Control Conference |
| ISBN: | 1479947288 , 3952426911 , 9781479947287 and 9783952426913 |
| Types: | Conference paper |
| DOI: | 10.1109/ECC.2014.6862441 |
| ORCIDs: | Andersen, Martin S. and Jørgensen, John Bagterp |
Benchmark testing Control systems Heuristic algorithms LQCP Optimization Power, Energy and Industry Applications Prediction algorithms Predictive control Riccati equations Riccati iteration procedure Robotics and Control Systems Signal Processing and Analysis Transportation Vectors convex optimal control problem convex programming direction method extended linear quadratic control problem highly structured quadratic program input-constrained model predictive control multipliers optimal control prediction horizon predictive control quadratic programming quadratic programming algorithms quadratic programs structure-exploiting interior-point method warm-start procedure
