Modelling studies in the Australian sugar industry

Systems simulation became a standard research and management technique in business, industry, and defence well before its use in agriculture became popular. Decisions made by agricultural managers in Australia are characterised by considerable uncertainty emanating from variable weather conditions, fluctuating prices, and institutional policy changes but in spite of well-developed theory about decision making under uncertainty being available from the 1960s, few practical applications have been seen of these methods. With the widespread availability of personal computers among agricultural scientists in the late 1980s and an upsurge in interest in modelling agricultural systems, many opportunities to develop aids to help agricultural managers with their decision-making were recognised. Two broad types of decision among those of interest to agricultural managers have been analysed in this thesis: the choice of appropriate production strategies on individual farms and resource management decisions for whole agricultural regions. Decisions of the first type include choice of cane varieties, and appropriate levels of production inputs such as irrigation water. Such decisions are made by individual producers and some examples from the sugar industry are described. The actions of whole groups of farmers acting together influences resource allocation at the regional level and again the sugar industry with its strong regional structure has been used to exemplify some possible applications of model-based decision-making techniques. The need for an adequate theoretical basis underpinning any aids that are developed to assist managers with both types of decision is discussed and the theory of decision making under uncertainty is extensively reviewed. In spite of its limitations, subjective expected utility theory remains the accepted theoretical basis for decision-making by individuals when outcomes are uncertain. It has been incorporated into a wide range of procedures that purport to provide information for decison-makers to enhance their ability to make better decisions when the outcomes from them are uncertain. A number of applications of these techniques have been reviewed. An attempt has been made in this thesis to extend the principles of decision-making by individuals under uncertainty to broader regional groups of decision-makers who act in concert, especially in regard to land utilisation. The collective behaviour of canegrowers in the choice of varieties has been shown to be consistent with mean-variance rules for decision-making. With the availability of a reliable crop simulation model for sugarcane, the implementation of stochastic efficiency rules for making these decisions is advocated. A mathematical programming model of regional canegrowing activities, based on information from a land use study and results generated by the simulation model, adequately described an optimal pattern of land use for the Moreton Mill area in southern Queensland when various assumptions were made about the area's participation in the continuing growth of the sugar industry