关键词: 生产计划和调度, 次年产量预测, 生产管理, 多元线性回归

Abstract: To overcome the difficulty of output prediction in production planning and scheduling,a novel output prediction approach based on dynamic-improved multiple linear regression model was proposed.The effecting factor of whole production cycle such as manufacturing resources,human input,manufacturing process and product rejection were taken as the modeling variables by multiple linear regression models in production management.Thus the prediction result of model was closer to the actual production.The initial multiple linear regression model was distinguished remarkably by backstepping method,and the low significance parameters for annual output were removed.On this basis,an improved multiple linear regression model was established.By analyzing the dynamic characteristics of this improved model,a dynamic improved model was obtained.This method was applied to the output prediction in the next year for an aviation manufacturing enterprise.The performance of this model was demonstrated by comparing the prediction output with actual output of the enterprise.

Key words: production planning and scheduling, output prediction, production management, multiple linear regression model