基于ANFIS和GA的多级分销链预测模型

为提高分销链企业的预测水平,优化分销链中各级企业库存管理,提出了基于自适应神经模糊推理系统(ANFIS)的单级预测模型和成本目标函数的多级预测模型。利用ANFIS模糊推理机制实现了其输入层与输出层间的非线性映射及该神经网络的信息存储和学习能力。从分销链整体成本优化的角度建立了多级预测模型并应用遗传算法(GA)进行求解。仿真结果表明,ANFIS与BP神经网络相比具有较高的准确性;在一定程度上改善了分销链中订货信息多级放大的现象。

Forecasting of Multi-echelon Distribution Chain based on Adaptive Neural-fuzzy Inference System and Genetic Algorithm

To improve the efficiency of forecasting demand and optimize stock management of all corporations in distribution chain,a single-echelon forecasting model based on the Adaptive Neural-fuzzy Inference System(ANFIS) and a multi-echelon forecasting model based on cost objective function are introduced.Fuzzy inference mechanism of ANFIS is used to fulfill a capacity of non-linear mapping between Input and output layer,and ability of information storing and learning of neural network.An optimization objection model is built to minimize total expected cost,and Genetic Algorithm(GA) is adopted to obtain the optimization.The results show that ANFIS is more accurate comparing with BP neural network.In the degree,multistage amplifier phenomenon of ordering variation is mitigated in distribution chain.