Research On Inventory Management Of Spare Parts For Urban Rail Transit Equipment

In recent years,China’s rail transit industry has developed rapidly.Maintenance spare parts play an important role in ensuring the long-term,safe and stable operation of the entire rail transit system.Scientific and efficient spare parts inventory management is of great significance for reducing rail transit operating costs.Most domestic rail transit spare parts are classified according to the economic indicators of spare parts and ABC classification method.The classification criteria of spare parts are single,which only considers the economic value of spare parts and ignores other attribute characteristics of spare parts,which is lack of rationality;In terms of demand forecasting,most of them adopt a single forecasting model,which can only meet the accuracy of demand forecasting for some spare parts,and the scope of application is limited;The unscientific classification of spare parts and the inaccurate demand forecast lead to the low efficiency of inventory control strategy management.On the basis of literature research,this paper sorts out the basic framework of spare parts inventory management research,analyzes the characteristics of rail transit spare parts management,and studies the spare parts classification method and demand forecasting method in view of the problems existing in rail transit spare parts classification and demand forecasting.The contents are as follows:Firstly,a multi index hierarchical clustering model of spare parts based on information entropy is proposed to solve the problem that there are many kinds and huge quantities of spare parts for rail transit equipment at present,and the attributes such as importance and unit price of spare parts are quite different and the classification criteria are single.Based on the analysis of the importance,economy and consumption of rail transit spare parts,the model determines several spare parts clustering indexes,and then uses information entropy to calculate the weight of each index.On this basis,the weighted Euclidean distance and average distance are selected as the calculation methods of sample distance and category distance of hierarchical clustering method,respectively,to realize the multi index clustering of rail transit spare parts.In the case study of rail transit spare parts classification,the classification results show that the multi index hierarchical clustering model of spare parts based on information entropy is significantly better than K-means clustering method and hierarchical clustering method in terms of contour coefficient SC and DB.Secondly,aiming at the problems of limited application scope and low accuracy of single model in spare parts demand forecasting,a combined forecasting model of spare parts demand based on artificial neural network is proposed.The model selects multiple single prediction models from the set of alternative models to predict the demand for spare parts,combines the predicted values of multiple single prediction models through neural networks,flexibly uses each single prediction model to capture information,expands the scope of application of the model,and realizes the accurate prediction of the demand for multiple types of spare parts for urban rail transit.In the case study of rail transit spare parts demand forecasting,the demand forecasting results show that the combined forecasting model of spare parts demand based on artificial neural network is superior to the single forecasting model such as DES and SVR and the simple arithmetic mean combined model in Mae,MAPE and RMSE.Finally,according to the spare parts classification and demand forecast results,appropriate inventory control strategies are formulated for various urban rail transit spare parts,which provides an important guarantee for the maintenance of rail transit equipment,provides a meaningful reference for improving the efficiency of rail transit spare parts inventory management,and meets the requirements of refined management of rail transit spare parts.