Chance‐constrained model predictive control

Stock management in hospital pharmacy using chance-constrained model predictive control

“…The rationale of this approach is to replace hard constraints with probabilistic constraints and the nominal cost function with its expected value in the MPC formulation [18], leading to a stochastic optimization problem. CC-MPC offers advantages as robustness, flexibility, low computational requirements, and the possibility of including the level of reliability associated with the constraints [19,20]. Furthermore, since CC-MPC takes into account the expected performance of the closed loop with probabilistic constraints instead of directly trying to assure robust constraint satisfaction, it avoids the conservatism present in other robust MPC techniques, e.g.…”

Model-based multi-parametric programming strategies towards the integration of design, control and operational optimization

“…[196] on-line min worst case quadratic cost, use of dynamic programming for closed-loop Schwarm and Nikolaou [319], Badgwell [15] on-line min nominal objective s.t. robustness quadratic/linear constraints.…”

An application of chance-constrained model predictive control to inventory management in Hospitalary Pharmacy

“…Combined with the standard MPC theory, they allow the designer to arise with a stochastic optimization problem behind the controller by replacing hard constraints (either of states or inputs) with probabilistic constraints and by replacing the nominal cost function with its expected value in the MPC formulation [4]. This stochastic approach, known as Chance-Constrained MPC (CC-MPC) demonstrates to be suitable for large-scale complex systems due to its inherent features such as robustness, flexibility, low computational requirements, and ability to include the level of reliability (or risk) associated with the constraints (which implies its a priori knowledge) [10], [5]. Thus, CC-MPC avoids the conservative nature of other MPC approaches taking into account the expected performance of the closed loop with proper constraint handling instead of directly trying to assure robust stability.…”