关键词: 电子商务, 双渠道闭环供应链网络, 牛鞭效应, 模型预测控制, 动态矩阵

Abstract: Aiming at the problem that Internet contained dual-channel forward logistics with Internet channel and traditional channel and dual-channel inverse logistics with return and remanufacturing,the system network dynamic model was established,and the general steps of model predictive control algorithm were presented.l2 norm measure methods of system bullwhip effect,forward bullwhip effect and remanufacturing bullwhip effect were given.With simulation examples,the smoothing effects of this method on production fluctuation,order fluctuation and inventory fluctuation of the system were discussed,and the control effects of this method on three kinds of bullwhip effect were researched.Further,the influence of channel preference on system bullwhip effect was discussed.The essence of the proposed algorithm was adjusted control variables such as order and production quantity through the rolling optimization and continuous implementations,which made the system have good dynamic performances in the premise of stability,and restrain the bullwhip effects.

Key words: e-commerce, dual-channel closed-loop supply chain networks, bullwhip effect, model predictive control, dynamic matrix