Application of Artificial Neural Network Supported by BP and Particle Swarm Optimization Algorithm for Evaluating the Criticality Class of Spare Parts

This paper presents artificial neural networks (ANNs) for the criticality class evaluating of spare parts in a power plant. Two learning methods are utilized in the ANNs, namely back propagation (BP) and BP-particle swarm optimization (BP-PSO). The reliability of the models is tested by comparing their classification ability with a hold-out sample and an external data set. The results show that both ANN models have high predictive accuracy. The results also indicate that the BP-PSO algorithm has better recognition rate than the BP algorithm. The proposed ANNs are successful in decreasing inventories holding costs significantly by modifying the unreasonable target service level setting which is confirmed by the corresponding criticality class of a spare part.