Discrete event sequencing as a traveling salesman problem

Abstract

Though many complex sequencing problems can be formulated as a Traveling Salesman Problem (TSP), the technique is not commonly employed for computational reasons. This research provides a novel perspective on the use of this methodology to the problem of steel product sequencing through the milling process of a large manufacturing facility. The purpose of this paper is to show how more traditional heuristic scheduling procedures may be restructured as TSP problems of manageable size and with significant improvement in attainment of production objectives. Modern steel production facilities are composed of many interrelated processes. The conventional approach to improving overall plant efficiency is to devise optimal operating strategies for each separate process. One such process, scheduling steel bars through the hot strip mill, is shown to lend itself well to optimization techniques when formulated and solved as the well-known Traveling Salesman Problem. Initial results on actual production schedules are shown to compare well with other scheduling techniques; improving efficiency by as much as 78% in total schedule penalty.