Abstract: In the general two-echelon inventory models, there exists a common assumption：the demand of every echelon is independent. But in fact, only the demand of the downstream node is independent, and the upstream node is correlated with the independent demand. Therefore, the collaborative ordering models of two-echelon inventory system based on hybrid demand are constructed. The different downstream nodes’ ordering can be transformed into a virtual node’s ordering. The upstream node takes a collaborative ordering. Thus, through constant adjustment of the downstream nodes’orders, it can get the optimal strategy to make system inventory cost minimum. Then the genetic algorithms of these models are given. The simulation shows that they are good ways to get the satisfied order policy.

Key words: two-echelon inventory, correlated demand, collaborative ordering, genetic algorithm

摘要： 在一般的二级库存优化模型均做如下假定：供应链的每一级节点面临的需求都是独立需求。但在实际中，只有最下游的节点面临的需求是独立需求，而其上游节点面临的需求都是与之相关的相关需求。因此，建立了基于混合需求的协同订货模型。将面临独立需求的最下游节点的订货合并为一个虚拟节点的订货策略，再对二级节点采用相关需求策略进行协同订货，从而构建使整体最优的模型。并给出了模型求解的遗传算法，最后对模型进行了实例演算。

关键词: 二级库存, 相关需求, 协同订货, 遗传算法