径向基函数网络的库存预测方法与仿真实验比较分析

Inventory Forecasting Method Based on Radial Basis Function Network

库存预测是企业编制物流系统存储、运输等各项业务活动计划的基础,是企业管理的一个重要环节。库存预测的精确与否,直接影响到企业计划的可行性,进而决定企业经营的成败。采用不规则需求预测法和本文提出的径向基函数网络需求预测法对不规则时间序列的需求预测问题进行了评价与分析。仿真实验表明,采用径向基函数网络需求预测法对不规则时间序列需求进行预测具有更强的针对性和有效性。该方法采用正交化方法和前向回归技术,计算量少,能够生成一个结构简单的神经网络预测模型。

Inventory forecasting method is the basis of enterprises to plan business operations such as the store and transportation of logistics system. Whether the inventory forecasting is precise or not directly affects the feasibiling of enterprise planning, which will result in success or failure of enterprise operations. The radial basis function network （RBFN） is presented for irregular demand time series. And the effectiveness of RBFN is demonstrated by sim^afion experiment. Adopting the orthogonal technology and forward regression, the RBFN method redues the amount of computation and yields a simpler network compared with traditional methods for irregular demand processes.