A dual gradient projection quadratic programming algorithm tailored for model predictive control

The objective of this work is to derive a QP algorithm tailored for MPC. More specifically, the primary target application is MPC for discrete-time hybrid systems. A desired property of the algorithm is that warm starts should be possible to perform efficiently. This property is very important for on-line linear MPC, and it is crucial in branch and bound for hybrid MPC. In this paper, a dual active set-like QP method was chosen because of its warm start properties. A drawback with classical active set methods is that they often require many iterations in order to find the active set in optimum. Gradient projection methods are methods known to be able to identify this active set very fast and such a method was therefore chosen in this work. The gradient projection method was applied to the dual QP problem and it was tailored for the MPC application. Results from numerical experiments indicate that the performance of the new algorithm is very good, both for linear MPC as well as for hybrid MPC. It is also noticed that the number of QP iterations is significantly reduced compared to classical active set methods.

[1]  Stephen J. Wright,et al.  Applying new optimization algorithms to more predictive control , 1996 .

[2]  Boris Polyak,et al.  Constrained minimization methods , 1966 .

[3]  V. A. Roshchin,et al.  Models and methods of solution of quadratic integer programming problems , 1987 .

[4]  J. J. Moré Gradient projection techniques for large-scale optimization problems , 1989, Proceedings of the 28th IEEE Conference on Decision and Control,.

[5]  G. Nemhauser,et al.  Integer Programming , 2020 .

[6]  Charles R. Johnson,et al.  Matrix analysis , 1985, Statistical Inference for Engineers and Data Scientists.

[7]  E. Polak,et al.  Constrained Minimization Problems in Finite-Dimensional Spaces , 1966 .

[8]  A. Goldstein Convex programming in Hilbert space , 1964 .

[9]  Stephen P. Boyd,et al.  ROBUST LINEAR PROGRAMMING AND OPTIMAL CONTROL , 2002 .

[10]  C. Panne,et al.  The Simplex and the Dual Method for Quadratic Programming , 1964 .

[11]  Alberto Bemporad,et al.  miqp. m: a Matlab function for solving Mixed Integer Quadratic Programs Version 1.02 - User Guide , 2000 .

[12]  D. Bertsekas On the Goldstein-Levitin-Polyak gradient projection method , 1974, CDC 1974.

[13]  Graham C. Goodwin,et al.  Constrained Control and Estimation: an Optimization Approach , 2004, IEEE Transactions on Automatic Control.

[14]  H. Jonson A Newton Method for Solving Non-Linear Optimal Control Problems with General Constraints , 1983 .

[15]  B. Lie,et al.  A comparison of implementation strategies for MPC , 2005 .

[16]  Donald Goldfarb,et al.  A numerically stable dual method for solving strictly convex quadratic programs , 1983, Math. Program..

[17]  Stephen P. Boyd,et al.  Convex Optimization , 2004, Algorithms and Theory of Computation Handbook.

[18]  Daniel Axehill,et al.  Integer Quadratic Programming for Control and Communication , 2008 .

[19]  Natashia Boland,et al.  A dual-active-set algorithm for positive semi-definite quadratic programming , 1997, Math. Program..

[20]  Stephen J. Wright,et al.  Numerical Optimization , 2018, Fundamental Statistical Inference.

[21]  L. Biegler,et al.  QPSchur: A dual, active-set, Schur-complement method for large-scale and structured convex quadratic programming , 2006 .

[22]  Alberto Bemporad,et al.  Control of systems integrating logic, dynamics, and constraints , 1999, Autom..

[23]  J. J. Moré,et al.  Algorithms for bound constrained quadratic programming problems , 1989 .

[24]  Sven Leyffer,et al.  Numerical Experience with Lower Bounds for MIQP Branch-And-Bound , 1998, SIAM J. Optim..

[25]  D. Bertsekas,et al.  CONVERGENCE OF A GRADIENT PROJECTION METHOD , 1982 .

[26]  Daniel Axehill,et al.  A Dual Gradient Projection Quadratic Programming Algorithm Tailored for Mixed Integer Predictive Control , 2008 .

[27]  M. Powell On the quadratic programming algorithm of Goldfarb and Idnani , 1985 .

[28]  D. Axehill,et al.  A Mixed Integer Dual Quadratic Programming Algorithm Tailored for MPC , 2006, Proceedings of the 45th IEEE Conference on Decision and Control.

[29]  Daniel Axehill,et al.  Applications of Integer Quadratic Programming in Control and Communication , 2005 .

[30]  P. Toint,et al.  Global convergence of a class of trust region algorithms for optimization with simple bounds , 1988 .

[31]  W. S. Dorn,et al.  Self-Dual Quadratic Programs , 1961 .

[32]  Daniel Axehill,et al.  A preprocessing algorithm for MIQP solvers with applications to MPC , 2004, 2004 43rd IEEE Conference on Decision and Control (CDC) (IEEE Cat. No.04CH37601).

[33]  Stephen J. Wright,et al.  Application of Interior-Point Methods to Model Predictive Control , 1998 .

[34]  D. Bertsekas Projected Newton methods for optimization problems with simple constraints , 1981, CDC 1981.

[35]  C. E. Lemke,et al.  A Method of Solution for Quadratic Programs , 1962 .