Research On The Model Of Controlling And Optimizing Supply Chain Multi-level Inventory

Under the economic environment of more and more sophisticated social division, the change of enterprise organizations and management styles has forced that supply chain stock control, which compares to the traditional inventory management, has many new features and requirements. Supply chain inventory optimization is one of important content of supply chain management, and it is effective method to strengthen the optimization of inventory control in order to increasing the systematic and agility of supply chain.In this thesis, first basic knowledge of inventory control, including supply chain structure, the composition of inventory cost and the strategy and method of controlling inventory, is introduced. And then the two stage and multi-echelon inventory models are established. At last the models are solved.At the premise permit of out of stock, the two stage inventory model, basing the supply chain composited by many supplies and one manufacturing plant, is builded to optimiate their order quantity. And under the condition of existing unconstrained minimum value in optimizition theory, the most optimiziting order quantity is derivated.The supply chain'structure of multi-echelon inventory control is composed of upstream suppliers and core manufacturer and downstream distribution centers and retailers, which there are complicated supply and demand relationships between. And the supply chain's shape likes a flatting sandglass.At the process of establishing multi-echelon inventory model, five aspects, including production, transportation, order response, inventory and cost are analysed. And the model arms for inventory costs, profits, retailers order response time and customer satisfaction and bases on supply chain network flow process of product delivery. The characteristics of the model can be sumed up with "three multiply", which are multiply targets, multiply periods and multiply products.The model is a multi-target and multi-cycle mixed integer programming, which it is difficult to solve by strict mathematical derivation or simplex method, so using genetic algorithms solves it. Genetic algorithm, which relies on the operators of selection, crossing and mutation are highly parallel search capabilities. The populations can improve themselves to reach better target.At last, through an example the paper verifies the feasibility and effectiveness of the models and algorithms.