Application of dynamic scheduling rules in maintenance planning and scheduling

Abstract

A maintenance planning and scheduling system often resembles that of a ‘job shop’ that is, the orders are one of a kind, and is characterized by having to schedule N orders through M or less tasks. The orders are of two types, e.g., (a) emergency—have to be done now, and (b ) non-emergency—can be delayed until later. In this type of ‘job shop’ the schedule becomes immediately out of date as soon as an emergency order is received. Consequently non-emergency orders are continually moved back in the schedule and forecasted completion dates are not met. Further if the orders entering the system exceed the normal available capacity, the backlog will continue to increase causing more disruption of schedules. The research, which is based on a large petrochemical plant, will deal with the above problems by (a) applying dynamic decision rules for day-to-day scheduling to ensure completion dates are met, (b) a method for controlling backlog, and (c) forecasting future load, and completion dates for orders. The results of the simulation experiments applied to the machine shop will be discussed.