Complex systems involve a number of interconnected entities, which collectively exhibit emergence and behaviour that cannot be explained by merely studying the individual entities. Agri-industrial systems, such as sugarcane production, are generally complex due to the presence of many autonomous stakeholders operating under diverse conditions, and may therefore contain varying perspectives and interests. The identification of problems and opportunities in such systems requires an approach that will, as far as possible, consider the entire system and how individual entities interact. Network analyses have the capacity to describe a complex system, depicting these interactions. In addition, graph theory approaches can help to identify key points in the system where there are opportunities for improvement. This paper presents a methodology to assist researchers to make sense of complex matters in an agri-industrial context. In the South African sugar industry, it can be argued that systemic inefficiencies in the supply chain reduce optimum performance. Research conducted in two large milling areas is used to develop and demonstrate the use of network approaches to analyse supply chains and identify opportunities for improvement. The research developed two types of map: system domain networks and theme networks, which are found to be appropriate for drawing a first set of conclusions concerning a relatively unfamiliar complex system. Although the paper focuses on sugarcane, there is significant scope to apply these techniques across a broader spectrum of agri-industrial sectors.

No abstract

The Eston Sugar Mill is the newest in the South African KwaZulu-Natal sugar belt. Like most other mills, it can be argued that there are inefficiencies in the supply chain due to systematic issues, which reduce optimum performance. It was alleged that mill processes are slowed, or stopped, on Sundays, Mondays, as well as some Tuesdays and Wednesdays, due to pay-weekends, because of the associated cutter absenteeism. This increases the length of the milling season (LOMS), increases milling costs and reduces the average cane quality for the season. Data on cane deliveries to the Eston Mill, over a period of five seasons, were analysed to study the magnitude of the problem. It was statistically verified that cane shortages occur immediately after payweekends and it was conservatively estimated that cutter absenteeism occurs between 25–29 days per season, which increases the LOMS by six to ten days. The associated cost of this problem equated to an average of US$159,500 (approximately EUR120,000) per milling season. In this paper, an alternative harvesting system scenario is suggested, assuming that mechanical harvesters be used after a pay-weekend, to mitigate the impacts of cutter shortages. However, the solution is calculated to be risky. When the cost of new equipment was considered, only two of the five seasons were able to justify the associated costs.