A Simulation Model for Harvest Operations under Stochastic Conditions

Abstract

The scheduled allocation of resources for large-scale harvesting operations is vulnerable to delays caused by variables of a stochastic nature such as rainfall and equipment breakdown. A computer simulation model was developed to test the merits of different harvesting strategies and decision rules on a sugar cane plantation. The model takes into account weather, hauling, equipment logistics, and equipment breakdowns. For a given man-machine harvest strategy and harvest schedule, it will simulate assigning harvest units (men and equipment) to fields, harvesting the cane, hauling the cane to mills, and milling. Submodels, based on a statistical analysis of historical data, generate values of simulated daily rain in each field, equipment breakdown, and repair times. Harvesting delays occur when the simulated rain reaches critical values. The hauling and milling capacity in the model depends to a great extent on both current and previous rainfall and is updated when required for new field conditions brought about by rain. The model keeps account of each cane-hauler round trip (some 50,000 trips a year) from field to mill, causes the haulers to queue when the loader or unloader is busy, and sends them to a repair shop when damaged. Simulated work hours and labor practices are also included in the model. For example, a submodel calls for overtime when progress indicates that the harvest may extend into the rainy season. The model gives periodic reports on the simulated production and information on capacity requirements for many of the tasks within the harvest operation.