Colloquium

Sugarcane harvest scheduling in Rwanda

Organisator Laboratory of Geo-information Science and Remote Sensing
Datum

vr 19 mei 2017 09:00 tot 09:30

Locatie Gaia, building number 101
Droevendaalsesteeg 3
101
6708 PB Wageningen
+31 317 48 16 00
Zaal/kamer 1

By Jorn Habes (The Netherlands)

Abstract
Sugarcane is an important commodity in the world. Rwanda is only a small player in the sugarcane business but the industry is of major importance on a national scale. Rwanda is a very poor country and the sugarcane industry supplies thousands of jobs. Kabuye Sugar Works, being the sole sugarcane processing plant in Rwanda, get its sugar from approximately 3000 fields along the Nyabarongo River. The sugar production in Rwanda faces multiple challenges both on the production side as well as the processing side. On the production side the maturity time (i.e. maximum sugar content) plays an important role. Examples of challenges faced by the Rwandese sugarcane industry are that the fields commonly flood, the harvesting is done manually, and the sugarcane processing factory has a minimum and maximum input. These are all examples of challenges that complicate harvesting in an optimal order. An optimisation model was created to schedule the harvest using available data. The model takes into account the constraints that complicate the harvest most. The goal of the model was to calculate the maximum possible revenue, and conclude whether harvest scheduling results in higher revenue than scheduling at random. This is done by maximising an objective function that is subject to these constraints, where simulated annealing was used as an optimiser. The model input consisted of growth models, field parameters such as height, ratoon cycle, sowing date, and eleven flooding scenarios which were created with both input data from the area as well as estimations based on expert opinion. The results showed that the model was capable of scheduling the harvest around the assumed optimal crop age, in this case 18 months. Each scenario achieved netto sugar yields of over 94% of the global optimum, showing the potential of sugarcane harvest scheduling in Rwanda.

Keywords: Constrained optimisation; Harvest scheduling; Kabuye Sugar Works; Rwanda sugarcane; Simulated annealing; Sugarcane harvesting.