This paper presents a methodology for deriving the best conic sector bounds to characterize a nonlinear multivariable plant. The plant nonlinearities are assumed to be Lipschitz within the plant's domain of operation. Our goal is to minimize the uncertain regions in parameter space required to represent the system nonlinearities over a specified compact domain. An optimization problem is formulated to find the best conic sector bounds. Although this problem is not convex, we make a number of observations and draw conclusions regarding the relationships between the variables, objective function and equality constraints. We use the Sequential Quadratic Programming method as implemented in MatLab's optimization toolbox to demonstrate the technique on a number of example 2-D nonlinearities.
|Number of pages||9|
|State||Published - 1 Dec 1994|
|Event||Proceedings of the 1994 International Mechanical Engineering Congress and Exposition - Chicago, IL, USA|
Duration: 6 Nov 1994 → 11 Nov 1994
|Conference||Proceedings of the 1994 International Mechanical Engineering Congress and Exposition|
|City||Chicago, IL, USA|
|Period||6/11/94 → 11/11/94|