A neural network based method for part demands prediction in auto aftermarket

Abstract

In the supply chain management of auto aftermarket, companies strive to manage inventory with low cost while maintaining a reasonable order fulfillment rate. To achieve this objective, a critical issue is to predict the demand for auto parts with high accuracy. Based on the factors relevant to auto aftermarket, this paper proposed an artificial neural network based method to forecast the demands of auto parts. The effectiveness of the proposed method is illustrated with a case study of an auto 4s shop in Shanghai.

Forecasting vehicle's spare parts price and demand

“…In a more developed country as Japan, according to Yu and Chen (2010), there are strict governmental regulations that demand vehicles inspections as a vehicle getting old which can be very expensive and tough, consequently Japanese customers usually buy a new vehicle every four years on average, they present Japanese Toyota company as a case study, the company have a considerable section for dealing with customer management, in order to keep customer information in a data base to keep in touch with them which increases customer satisfaction and loyalty to the company, however, vehicle customers and companies in other countries especially the developing countries do not follow the same procedure, as they are more often replacing vehicle parts with SPs. After the end of warranty duration, Corrêa et al (2007) find that the customers have options to maintain their vehicles either by branded vehicle dealers which provide vehicle maintenance during warranty or by independent garages, a comparison was held between the two options and the results were analysed.…”

“…Whereas Li and Kuo (2008) develop an advance forecasting SPs demand method; the fuzzy neural network (FNN), the parts are automobiles spares which are held in a main warehouse. Yu and Chen (2010) suggests a new forecasting demand method in automotive SPs; the artificial neural network, the objective of this study is to develop services and decrease the operational cost thus improving supply chain management. Finally, Wattanarat et al (2010) is the only found study to combine demand with price forecasting in one study; it focuses on increasing the precision of forecasting demand and price by comparing a real option method which is related to financial studies based on geometric Brownian motion and investments with the autoregressive integrated moving average (ARIMA) method using the MATLAB to big data analysis, this research uses realistic data of die casting company and copper price.…”

Auto parts demand forecasting based on nonnegative variable weight combination model in auto aftermarket

“…In recent years, some researches have been done in the domain of how to forecast auto parts demand for auto aftermarket. Most of the present forecasting models have used an individual model, such as a neural network based method [1] , Regression-Bayesian-BPNN model [2] , ARMA model [3] , etc. Besides forecasting based on historic sales data, some researchers have proposed demand forecasting model based on dynamic fault law [4] .…”