Optimal scheduling of a repair shipyard based on Supervisory Control Theory

Abstract

Supervisory Control Theory (SCT) allows for the automatic synthesis of non-blocking supervisors capable of ensuring all and only those sequences of events that meet the specifications for discrete event systems. For a job shop problem, it's possible to obtain an optimal supervisor capable of satisfying resource, production routine and deadline specifications, as well as of providing all viable scheduling solutions, by using a timed SCT extension. This paper investigates the feasibility of applying SCT to production planning and logistics modeling - applications rendered possible through a method for synthesizing schedules via an innovative approach to resource modeling; through the incremental synthesis of production routines and deadlines; and through a bisection method for minimizing total production time and job-specific lead times. The method is applied to the scheduling of activities at a repair shipyard, namely, the activities of the five main resources involved in the execution of ten job orders. The paper concludes with the presentation of a computational tool developed to integrate production planning and a tool for automatic supervisory synthesis so as to facilitate handling of the interface by users not familiar with SCT.