Accelerated and Adaptive Modifier-Adaptation Schemes for Unconstrained Uncertain Systems. Issue 1 (July 2017)

• Record Type: Journal Article
• Authors: Schneider, René; Milosavljevic, Predrag; Bonvin, Dominique
• Is Part Of: IFAC-PapersOnLine. Volume 50:Issue 1(2017)
• Page Start: 4684
• Page End: 4689
• Publication Date: 2017-07
• Subjects: Modifier adaptation -- real-time optimization -- proximal gradient -- model uncertainty -- guaranteed convergence -- accelerated convergence; Automatic control -- Periodicals; 629.805
• DOI: 10.1016/j.ifacol.2017.08.719
• Languages: English
• ISSNs: 2405-8963

Abstract

Abstract: The steady-state performance of a parametrically or structurally uncertain system can be optimized using iterative real-time optimization methods such as modifier adaptation (MA). Here, we extend a recently proposed MA scheme in two important and novel directions. First, we accelerate its convergence, i.e., we reduce the number of potentially time-consuming and suboptimal transitions to intermediate steady states by appropriate filtering. Second, we propose an adaptation strategy to reduce conservatism related to the unknown curvature of the system's steady-state performance curve. Moreover, we combine these two innovations and demonstrate their benefits on two numerical examples.