ARIMA models to forecast demand in fresh supply chains

This paper presents the application of autoregressive integrated moving average (ARIMA) models to forecast the demand of fresh produce (fruits and vegetables) on a daily basis. Models were built using 25 months sales data of onion from Ahmedabad market in India. Results show that the model can be used to forecast the demand with mean absolute percentage error (MAPE) of 43.14%. This error is within the acceptable limit for fruits and vegetable markets with highly fluctuating demand pattern. The model was validated taking sales data for the same commodity from a different vegetable market. The proposed forecasting model can be used to assist the farmers in determining the volume of daily harvesting for fruits and vegetables.