Safeguarded optimal policy learning for a smart discrete manufacturing plant. Issue 2 (2022)

• Record Type: Journal Article
• Authors: Boffadossi, Roberto; Bonassi, Fabio; Fagiano, Lorenzo; Scattolini, Riccardo; Cataldo, Andrea
• Is Part Of: IFAC-PapersOnLine. Volume 55:Issue 2(2022)
• Page Start: 396
• Page End: 401
• Publication Date: 2022
• Subjects: Advanced Manufacturing -- Nonlinear Model Predictive Control -- Machine Learning for Control -- Safe Learning -- Neural Networks; Automatic control -- Periodicals; 629.805
• DOI: 10.1016/j.ifacol.2022.04.226
• Languages: English
• ISSNs: 2405-8963

Abstract

Abstract: An approach to safely learn and deploy, at fast rate, a given optimization-based controller for the routing problem in smart manufacturing is presented. The considered application features a large number of integer decision variables, combined with nonlinear dynamics, temporal-logic constraints, and hard safety constraints. The approach employs a neural network as feedback controller, trained using a data-set of state-input pairs collected with the optimization-based controller. A safeguard mechanism checks whether the input computed by the neural network is feasible or not, in the latter case the optimization-based controller is called. Results on a high-fidelity simulation suite indicate a strong decrease of average computational time combined with a negligible loss of plant performance.