The Optimization of Paper Machine Scheduling

Abstract

A method for determining schedules in a paper mill is described and applied to real production data. The problem is to allocate products to machines and determine their sequence so as to minimize total costs. The costs include production losses through set-ups, stock holding and penalties for delayed delivery. The method is based on a permutation representation of the production schedule and the determination of a locally optimal solution. A permutation is locally optimal if it cannot be improved by transferring groups of products to a different place in the schedule. The method has been programmed and the results show a considerable improvement over alternative manual procedures. The paper has been written in two parts, the first dealing with the development of the method and the second discussing its application to a specific problem.