The Impact of Requirement Splitting on the Efficiency of Supervisory Control Synthesis

Abstract

Supervisory control theory provides means to synthesize supervisors for a cyber-physical system based on models of the uncontrolled system components and models of the control requirements. Although several synthesis procedures have been proposed and automated, obtaining correct and useful models of industrial-size applications that are needed as their input remains a challenge. We show that the efficiency of supervisor synthesis techniques tends to increase significantly if a single large requirement is split into a set of smaller requirements. A theoretical underpinning is provided for showing the strength of this modeling guideline. Moreover, several examples from the literature as well as some real-life case studies are included for illustration.