Developing and implementing optimised sugarcane harvest schedules through participatory research

Abstract

The Australian sugar industry saw opportunities for increasing productivity and hence whole- of-industry profitability through optimising the harvest date of sugarcane, accounting for geographical and crop differences in cane yield and the sugar content of cane for different harvest dates throughout the harvesting season. Research scientists engaged in participatory research with 3 case-study mill regions to construct the models needed to produce these optimised harvest schedules. Average potential gains of up to AU$119/ha at a sugar price of AU$250/t were shown and the case study regions were keen to pilot the schedules. This paper focusses on the development of a pathway to pilot implementation and evaluation through collaboration with growers, harvester contractors, and millers collectively. We developed: strategies to overcome implementation barriers; the design of the pilot scheme; software tools; and an evaluation strategy of pilot study results to provide proof-of-concept and encourage further uptake. With the development of this pathway, pilot implementation took place during the 2000 and 2001 harvest seasons for about 200 farms. Action learning methodologies were applied to improve the scheme for the 2001 season. Most growers and millers who followed the optimised schedules closely, achieved gains of up to AU$200/ha, with the Maryborough Sugar Factory estimating a gain of AU$34 000 for their crop. Although optimised harvest schedules were implemented by only a small percentage of growers in each of the case study regions, there is growing adoption throughout the sugar industry.

Lack of consideration for end-users during the design of agronomic models. A review

“…Girard and Hubert (1999) assume that "Formalising the knowledge to be used in gaining this understanding is therefore of crucial importance in building such tools". Higgins et al (2004) explain that "Research scientists engaged in participatory research with 3 casestudy mill regions to construct the models needed to produce these optimised harvest schedules".…”

Programação e sequenciamento das frentes de colheita de cana-de-açúcar: modelo e métodos de solução para problemas de grande porte

“…In order to address the actual sugarcane harvesting planning, Higgins et al (2004b), Higgins & Laredo (2006), as well as Jena & Poggi (2013) proposed methods that aggregate harvest blocks in order to reduce the number of variables involved. In parallel, heuristic methods based on MIP, such as relax-and-fix, presented by Pochet & Wolsey (2006) and fix-and-optimize, proposed by Helber & Sahling (2010), have been widely used to solve the large scale General Lotsizing and Scheduling Problem for Parallel Production Lines, as shown for example in Beraldi et al (2008), Ferreira et al (2009Ferreira et al ( , 2012, Toso et al (2009) and Helber & Sahling (2010).…”