Optimising the transportation of avocados from farm to packhouse using Bayesian networks

Abstract

Unnecessary losses occur due to the postharvest transportation of avocados from farm to packhouse. Damage done to avocados may only become visible during later stages of the fruit's ripening, making it difficult to detect damage during the early stages of the supply chain. This study looks at where improvements can be made, specifically identifying the hazards occurring between harvest and the packhouse, by means of a Bayesian network. The network was populated using data collected from the farm using civiltronics, investigating hazards by consulting literature and using expert elicitation to better understand the identified hazards and their effect. The study concluded that the most probable cause of damage results from the overall tree condition, the delay in transportation, picking techniques and unloading procedures at the packhouse. The Bayesian network is a powerful tool which can be updated with evidence from the farmers once improvements are made to reassess the network.