The Coordinated Decision Of Multistage-inventory In Retail Chain Based On Lean Logistics Supply Chain Management

With the high-speed economic development over30years of china’s reform andopening up, the comprehensive strength of China has been improved a lot. Chainsupermarket, as an important member of the retail chain, also experienced a rapiddevelopment. China chain management association published a list of top100, it saidretail-chain has been the main enterprises, account for more than half during the year of2005-2010. We can see that the retail-chain’s growth trend is very obvious.Retail-chain enterprise is a typical multi-stage supply chain system, and supplychain management is a very popular management pattern. The core idea of LeanLogistics is about a continuous improvement system of ideology which focus on how toavoid all the waste. It replies customers as quickly as possible with low cost and highquality, and thus enhances the enterprise’s competitive ability. Inventory management isthe indispensable and important part in the process of production and management,which is an important link to achieve the added value. In Lean Logistics Supply ChainManagement, inventory is the biggest obstacle. The discretion of the inventory not onlyaffects the single enterprise, it also limits the performance of the whole supply chain.Along with the increase customers’ demand diversity and uncertainty, how to set up theinventory is also the main question for the managers who try to improve and optimize thewhole supply chain. To sum up, looking for a best inventory strategy, reducing inventorylevels throughout the supply chain and decreaseing the node inventory cost are the mostimportant problems of the research to the lean logistics supply chain.With this background, this thesis is to analysis the retail-chain multi-stage inventory.the complexity of multi-stage inventory is that the stock holders in the supply chain arein different levels and different link of the chain. By all kinds of uncertainty factors, thecomplex is much higher than single node inventory problem. For the moment, singlesupply chain inventory problen has had a lot theories and practice results, but for twolayer or above, the research has achieved very little. This thesis, first of all, makesclassification of all kinds of commodities in supermarket according to sales, inventorycharacteristics and chooses one of the common items as the research object. Second, thisthesis forecasts the inventory terminal demand with wavelet neural network and hasmade some achievements. But turned out that simply wavelet neural network has manyflaws such as long forecast time, easy to produce the local minimum. So on the basis of the original wavelet network, this paper introduces a variable step factor and momentumfactor for improvement, and make the system achieve the ideal level. Third, this paperuses multi-agent modeles to optimize retail chain multilevel inventory system. It seesevery branch as son agents, distribution centers as father agents. Finally, it selectsNETLOGO as the multi-agent simulation platform, and compare with other methods’simulation results. Fourth, it uses the interactive multi-agent modele to optimizeinteractive systems based on JAVA-JADE and solve the retail chain widespreadproblems of disorderly delivery delay and delivery of disorder caused by interactivemixed.In this paper, the wavelet neural network, the decision of multilevel intelligentcountries combining with inventory modeling and simulation and the JAVA-JADEmulti-agent interaction system has been formed to a whole multilevel inventorycoordinated decision system from upstream to downstreem. This study is not only ainnovation to supply chain inventory, but also a completion to the retail chain multilevelinventory research. Whether the theoretical research or practices are all have certainvalues and meanings.