Article ID: | iaor20071843 |
Country: | United Kingdom |
Volume: | 57 |
Issue: | 5 |
Start Page Number: | 490 |
End Page Number: | 498 |
Publication Date: | May 2006 |
Journal: | Journal of the Operational Research Society |
Authors: | Higgins A.J., Harrison A., Beashel G |
Keywords: | scheduling, programming: integer |
The Australian sugar industry exports about 90% of its five million tonne annual sugar production on the open market to customers in several countries, who request various brands of raw sugar. In planning at this interface of the sugar supply chain, an objective is to schedule the brands of sugar each mill produces throughout the annual harvest season, along with the ports that each ship loads at, to minimize total costs of sugar production and shipping. The complexity of such planning for the sugar industry has been overcome through the development and solution of a mixed-integer programming model to perform the task. The model is used as an annual planning tool to obtain a base schedule, as well as a rescheduling tool to revise the plan during the year as mill production rates and shipments change. Using the Australian sugar industry as a case study, this paper focuses on the mathematical model and solution using two known meta-heuristics based on local search. Using the 2002/2003 and 2003/2004 financial years, a comparison is made between the schedules produced in practice using manual methods and those using the model, which show a total potential cost savings of up to AU$4.0 M for the 2 years.