Research on improving prediction of demand for common components through aggregation effect—Simulation based on linear trend of ols method

摘要

The shift the prediction object from the final product to the common component makes better use of the aggregation effect. In this study, the mutually independent demands for all the products with linear increasing trend is generated by the method of Monte Carlo Stochastic Simulation. Using the prediction method of ordinary least square, the comparison is made with two prediction results. One of which is to predict final products demand, then the total demand for common components of all the products is calculated based on the Bill of Materials, and the other is to predict the demand of common components directly. The following conclusion is drawn that the direct prediction of the demand for common components could take advantage of the aggregation effect more sufficiently. This study also discusses the influences of correlations of the product demands, the fluctuation degree, and the number of the common components, etc. on the aggregation effect.