Prediction of tractor repair and maintenance costs using Artificial Neural Network

The prediction of repair and maintenance costs has significant impacts on proper economical decisions making of machinery managers, such as machine’s replacement and substitution. In this article the potential of Artificial Neural Network (ANN) technique has evaluated as an alternative method for the prediction of machinery (specifically tractor) repair and maintenance costs. The study was conducted using empirical data on 60 two-wheel drive tractors from Astan Ghodse Razavi agro-industry in Iran. Optimal parameters for the network were selected via a trial and error procedure on the available data. In this paper, the performance of Basic Back-propagation (BB) training algorithm was also compared with Back-propagation with Declining Learning-Rate Factor algorithm (BDLRF). It was found that BDLRF has a better performance for the prediction of tractor’s costs. The prediction of repair and maintenance cost components of tractors with a single network produced a better result than using separate networks for prediction of each cost component. It has been concluded that ANN represents a promising tool for predicting repair and maintenance costs.