基于权值直接确定神经网络的采购提前期模型

摘要

精确采购，尽可能地降低库存成本是大多数企业追求的目标之一。然而准时采购受到库存量、采购处理等众多因素影响，这些因素之间彼此影响，因此对采购提前期的预测很难给出一个精确的数学模型。人工神经网络具有很强的非线性映射能力，可以很大程度地解决非线性预测问题，对此提出一种权值直接确定神经网络的采购提前期模型，仿真效果达到预期水平。%Precise purchase to reduce the cost for inventory as far as possible is one of the goals that most enterprises pursuit. However, just in time purchasing is affected by many factors, such as inventory and purchase process. These factors influence each other, so it is difficult to make a precise mathematical mod⁃el to forecast the purchasing lead-time. ANN has strong ability of nonlinear mapping and can solve the nonlinear mapping fore⁃casting problem well. A purchasing lead-time model based on weights direct determination neural network is given in the arti⁃cle and the simulation effect achieves the expected level.