Suitability of Redesigned Digital Control Systems Having an Interval Plant via an Evolutionary Approach

Abstract

In this paper, a quantitative index is proposed to address the performance evaluation and design issues in the digital redesign of continuous-time interval systems. From the perspective of signal energy, a worst-case energy resemblance index (WERI), defined as the ratio of the worst-case continuous signal energy (WCSE) of the continuous-time interval system over the worst-case discrete sequence energy (WDSE) of the redesigned digital system, is established for evaluating the closeness of the system performance between the redesigned digital control system and its continuous-time counterpart. Based on the WERI, performance of the redesigned digital systems can be evaluated for different discretization methods at different sampling times. It is found that no discretization method outperforms the others for all sampling times. Because of serious nonlinearities and nonconvexity involved, the determination of WCSE and WDSE is first formulated as an optimization problem and subsequently solved via an evolutionary algorithm. To guarantee stability of the redesigned digital system, the largest sampling time allowed is also evolutionarily determined to establish a sampling-time constraint under which robust Schur stability of the redesigned digital system can be ensured. For design purposes, sampling time required can be determined according to the user-specified WERI, which serves as a performance specification for fine tuning the performance of the redesigned digital control system.