Study On Slow Moving Spare Parts Supply Effectiveness-Oriented Demand Forecasting

Slow-moving spare parts are the main form of spare parts for enterprises,and are also the main source of spare parts inventory costs.Demand forecasting is an important solution to reduce the inventory of slow-moving spare parts.However,due to its serious data discontinuity,complex enviromental changes with its long life cycle,it is difficult to achieve the desired effect using conventional prediction techniques.This paper takes the slow-moving spare parts demand forecasting problem as the research object,deeply analyzes the slow spare parts demand characteristics,and defines the effective concept of demand forecasting based on the clear demand forecasting profit point,and studies the external resource configuration method.Under the condition of established supply capacity,the impact of time-effective and quantity-effective forecasting methods on inventory is analyzed for slow-flowing spare parts.Based on this,time-effectiveness is defined and a time-effectiveness-based propulsion-type slow-moving spare parts demand forecasting method is established to ensure that the forecast results played an effective role in the spare parts supply system.The main research contents include:(1)The division of spare parts supply status based on demand forecasting effectiveness analysis is proposed.For the problem of forecasting pressure caused by the complexity of the spare parts,a supply state partition map is presented at the perspective of demand forecast analysis.It has realized the reduction for demand forecasting,and pointed out the technical route of the improvement of demand prediction ability in different supply states.Based on the analysis of spare parts demand and forecast characteristics,combined with the requirements of demand forecast in time dimension and quantity dimension,the validity of spare parts demand forecast is defined,meanwhile the spare parts supply state is divided into three status zone according to the predictive effectiveness of the time dimension,and an estimation method for the critical point of predictive effective in time dimension are discussed.The zoning process of the supply status of spare parts can effectively assess the status of demand forecasting and find an effective improvement route for improving the utility of spare parts demand forecasting.This method can provide enterprises with an effective technical route to improve the ability of forecasting spare parts,and ensure that enterprises can realize the basic tools of forecasting effects with limited resources.(2)An external resource allocation method for advancing the implementation of slow spare parts prediction utility is proposed.The utility of spare parts prediction can only be achieved when the external supply source reaches a certain capacity,the supply capacity of external sources is mainly derived from the supplier's supply capacity.Therefore,in order to ensure that the external resource allocation capability meets the basic requirements for predictive utility realization,an external supplier evaluation method for the effectiveness of spare parts demand forecasting is proposed.In order to achieve the objective of advancing validity of forecasting,the presented method employ Decision Tree and Support Vector Machine for classifying to spare parts,and get the supplier's category that is divided by category of spare parts,and then adopt the strategy of differentiated management for the spare parts suppliers,meanwhile some supplier management strategies suitable for the production of the actual situation are proposed for the single class suppliers and complex class suppliers.The feasibility and effectiveness of this method are applied and fully verified by a practical case,and forecasting the guiding role of work on supply work is effectively improved.(3)A propulsion prediction method for improving the internal prediction capability of slow spare parts is proposed.For the three key problems of affecting the prediction accuracy of slow moving spare parts,a propulsion prediction method is proposed based on the comparison of operating conditions and performance state double similarity.? For the spare parts demand forecast is less accurate caused by the special form of historical data,a historical data reconstruction model based on the work station of in-service parts has been proposed.The presented model transforms the common intermittent sequence prediction of demand based on fixed time into demand time continuous sequence prediction based on a fixed number,meanwhile the original discontinuous time series prediction problem is transformed into a more accurate continuous time series prediction problem.? On the basis of data reconstruction,for the problem of poor accuracy of slow moving spare parts caused by significant differences in historical sample conditions,the condition pattern recognition based on RS and DT,and the reference sample extraction method based on the comparison of the similarity of working conditions are proposed.Based on the key factor identification,a similarly sample set is obtained as the forecast sample set,by comparing the similarity of working conditions.? For the long-range prediction problem of slow moving spare parts,a propulsion prediction model based on the similarity of spare parts performance states,PABDP(Performance Assessment Based Demand Prediction)is proposed.A variable step size iterative prediction strategy based on the procurement cycle is used for the model,the failure warning time is obtained according to the current deterioration state curve of the in-service parts and the position where the similarity degree is the greatest in each sample state deterioration curve;the demand interval of the parts under the current operating condition is predicted by all the failure warning time,and in accordance with its relationship with the warning criteria,the decision result that issued a procurement warning or change the forecast interval for iterative prediction is outputted by the model,hereto,the validity forecast of slow moving spare parts is realized.In order to solve the problem of improving the guidance of demand forecasting in slow moving spare parts supply,a set of technical route for slow moving spare parts demand forecasting to promote effective forecasting is constructed,meanwhile,a series of research methods to improve the ability of slow moving spare parts demand forecasting under different supply states are proposed.When the forecasting time results is lose efficacy,the effectiveness of forecasting is enhanced by properly allocating external resources.When the forecasting time results is efficacy,through the historical data reconstruction based on the station,and the double similarity analysis of the external working conditions and the performance status of the parts,the problems affecting the internal forecasting ability of the spare parts that the data discontinuity,working condition impact and long-range prediction are solved respectively.This study provides a new method and idea for the theoretical research and practice of slow moving spare parts demand forecasting.