1 June 2018 in Computers & Chemical Engineering
10 May 2017 in International Journal of Control
This article investigates model predictive control (MPC) of linear systems subject to arbitrary (possibly unbounded) stochastic disturbances. An MPC approach is presented to account for hard input constraints and joint state chance constraints in the presence of unbounded additive disturbances. The Cantelli–Chebyshev inequality is used in combination with risk allocation to obtain computationally tractable but accurate surrogates for the joint state chance constraints when only the mean and variance of the arbitrary disturbance distributions are known. An algorithm is presented for determining the optimal feedback gain and optimal risk allocation by iteratively solving a series of convex programs. The proposed stochastic MPC approach is demonstrated on a continuous acetone–butanol–ethanol fermentation process, which is used in the production of biofuels.
1 June 2015 in Computers & Chemical Engineering
22 February 2015 in ACS Nano
12 March 2020 in IEEE Control Systems Letters
The real-time implementation of closed-loop robust model predictive control (MPC) schemes is an important challenge for fast systems, as their solution complexity depends strongly on the system size, control policy parametrization, and prediction horizon. We look to address this problem by approximating the implicitly-defined MPC controller using deep learning. Although the resulting neural network approximation has a small memory footprint and can be quickly computed, it does not guarantee robust constraint satisfaction or stability. We propose a novel projection-based strategy that is capable of providing a certificate of robust feasibility and input-to-state stability in real-time. We also show how this projection operator can be formulated as a parametric quadratic program that is solvable offline. The advantages of the proposed approach are demonstrated on a benchmark case study.
1 June 2017 in Organic Process Research & Development