作者: Mircea-Bogdan Radac , Radu-Emil Precup , Emil M. Petriu , Stefan Preitl
关键词: Mathematical optimization 、 Control theory 、 Computer science 、 Nonlinear system 、 Aerodynamics 、 Angular displacement 、 Optimization problem 、 Constrained optimization 、 Control theory 、 Data-driven 、 Search algorithm
摘要: This paper presents a new iterative data-driven algorithm (IDDA) for the experiment-based tuning of controllers nonlinear systems. The proposed IDDA solves optimization problems processes while using linear accounting operational constraints and employing quadratic penalty function approach. search employs first-order gradient information obtained from neural-network-based process models to reduce number experiments needed run on real-world processes. A controller angular position control aerodynamic system is used as an experimental case study validate IDDA.