Deep learning-based embedded mixed-integer model predictive control

We suggest that using deep learning networks to learn model predictive controllers is a powerful alternative to online optimization, especially when the underlying problems are complex, as in the case of mixed-integer quadratic programs. The use of deep learning has two important advantages compared to classical shallow networks regarding its embedded implementation. A better function approximation can be achieved with the same number of neurons and less weights are necessary due to the use of more, but smaller, layers. This reduces significantly the memory footprint of the necessary code for its embedded implementation. As with shallow networks, deep neural networks are extremely easy to implement and to deploy on embedded platforms. The potential of the approach is illustrated with simulation results of an energy management system in a smart building, including the implementation of the proposed controller using a low-cost microcontroller.

[1]  Rolf Findeisen,et al.  Efficient stochastic model predictive control based on polynomial chaos expansions for embedded applications , 2015, 2015 54th IEEE Conference on Decision and Control (CDC).

[2]  Daniel E. Quevedo,et al.  Predictive control, embedded cyberphysical systems and systems of systems - A perspective , 2016, Annu. Rev. Control..

[3]  Bart De Schutter,et al.  Accelerated gradient methods and dual decomposition in distributed model predictive control , 2013, Autom..

[4]  Alberto Bemporad,et al.  The explicit linear quadratic regulator for constrained systems , 2003, Autom..

[5]  Rolf Findeisen,et al.  μAO-MPC: A free code generation tool for embedded real-time linear model predictive control , 2013, 2013 American Control Conference.

[6]  Benjamin Müller,et al.  The SCIP Optimization Suite 5.0 , 2017, 2112.08872.

[7]  Alberto Bemporad,et al.  A survey on explicit model predictive control , 2009 .

[8]  Jimmy Ba,et al.  Adam: A Method for Stochastic Optimization , 2014, ICLR.

[9]  Rolf Findeisen,et al.  A fast gradient method for embedded linear predictive control , 2011 .

[10]  Denis Navarro,et al.  Optimized FPGA Implementation of Model Predictive Control for Embedded Systems Using High-Level Synthesis Tool , 2018, IEEE Transactions on Industrial Informatics.

[11]  Vihangkumar V. Naik,et al.  Embedded Mixed-Integer Quadratic Optimization using Accelerated Dual Gradient Projection , 2017 .

[12]  Sebastian Engell,et al.  Multi-stage nonlinear model predictive control applied to a semi-batch polymerization reactor under uncertainty , 2013 .

[13]  Stephen P. Boyd,et al.  CVXGEN: a code generator for embedded convex optimization , 2011, Optimization and Engineering.

[14]  Demis Hassabis,et al.  Mastering the game of Go with deep neural networks and tree search , 2016, Nature.

[15]  Stephen P. Boyd,et al.  Distributed Optimization and Statistical Learning via the Alternating Direction Method of Multipliers , 2011, Found. Trends Mach. Learn..

[16]  Béla Lantos,et al.  Explicit MPC-Based RBF Neural Network Controller Design With Discrete-Time Actual Kalman Filter for Semiactive Suspension , 2015, IEEE Transactions on Control Systems Technology.

[17]  Stephen P. Boyd,et al.  A simple effective heuristic for embedded mixed-integer quadratic programming , 2015, 2016 American Control Conference (ACC).

[18]  Martín Abadi,et al.  TensorFlow: Large-Scale Machine Learning on Heterogeneous Distributed Systems , 2016, ArXiv.

[19]  Ohad Shamir,et al.  Depth-Width Tradeoffs in Approximating Natural Functions with Neural Networks , 2016, ICML.

[20]  Manfred Morari,et al.  A tractable approximation of chance constrained stochastic MPC based on affine disturbance feedback , 2008, 2008 47th IEEE Conference on Decision and Control.