Cloud-based model predictive control with variable horizon
Per Skarin, Johan Eker, Karl-Erik Årzén
Abstract
A novel method using the cloud to implement a variable horizon model predictive controller is presented. In case of sudden long delays and downtime, a graceful degradation is used. Robust, best effort strategies allow industrial grade use of the powerful, efficient, and quickly improving cloud ecosystems. The variable horizon strategy finds use in, for example, non-linear control problems, and the proposed method can be generalized to implement robust and scalable controllers that benefit from cloud technology. We show results from two horizon selection strategies, service degradation and connectivity issues.
Topics & Concepts
DowntimeCloud computingModel predictive controlHorizonComputer scienceVariable (mathematics)Time horizonScalabilityController (irrigation)Control theory (sociology)Mathematical optimizationControl (management)Artificial intelligenceMathematicsAgronomyMathematical analysisDatabaseOperating systemGeometryBiologyAdvanced Control Systems OptimizationFault Detection and Control SystemsProcess Optimization and Integration