A novel prediction model for aircraft spare part intermittent demand in aviation transportation logistics using multi-components accumulation and high resolution analysis

Abstract

High effective management of civil aircraft spare parts is of significant importance for the economical operation of aircrafts. However, the stochastic characteristics of the aircraft spare parts make it difficult to find a reliable rule to precisely predict the future demand. In order to address this issue, this work presents a novel multi-components accumulation and high resolution analysis (MCAHR) method to improve the forecasting performance of aircraft spare parts. The MCAHR takes the advantages of high resolution of the wavelet transform to analyze the time series of spare part intermittent demand. The original time series were decomposed into several sub-bands along with the time axis of the wavelet. Then particle swarm optimized fuzzy neural networks were established for each sub-band to intelligently mine their intrinsic features. Accurate prediction result was hence obtained by the accumulation of the outputs of all fuzzy neural networks. Experimental tests using the historical data of A320 civil aircrafts were carried out in this work to evaluate the proposed MCAHR method. The analysis results have demonstrated a high efficiency of the MCAHR method and that its prediction performance is superior to existing methods. Hence, the proposed MCAHR method has practical importance in the civil aircraft spare part intermittent demand prediction and will provide a significant economic benefit to the industry through reasonable management of aircraft spare parts.