Semi-global stability analysis of a discrete-time extremum-seeking scheme using LDI methods

This paper presents and analyses a new formulation of extremum-seeking control for discrete-time systems. Assuming a static, strongly convex SISO plant, a novel method of analysis using Linear Difference Inclusions (LDI) demonstrates semi-global stability of the control scheme. Furthermore, given limited information about the plant mapping, it is shown how the LDI system description allows a conservative upper bound on the adaptation gain to be evaluated, improving convergence speed.