Research Of Supply Chain Pre-Alert With Cognitive Map Based On Partical Swarm Optimization

With the globalization process, the market competition is not only the competition between one enterprise and another,but also the competition between supply chains. Supply chain is a complex network,includes the collaboration among market, sales, production, product design, purchase, logistics, finance and information techology. Any problem in any link will influences the other links,even cause the breakdown of the whole supply chain. Supply chain network risks originate from many elements,and the risks have apparent or inapparent relationships between them. Supply chain is not a steady network but is changing all the time, all these factors make the study of supply chain risks a big challenge.This paper deals with supply chain risks, takes advantages of cognitive map and particle swarm optimization. I came up with the idea that simulate the supply chain network with cognitive map, optimize the map with particle swarm optimization method. This method simulated the runtime situation of supply chain network, solve the problem of the previous models—lack of self-study ability, and provide us a realizable way to study the supply chain pre-alert.Firstly,simulate the supply chain network with cognitive map. Cognitive map can model the elements and the relationships between them, especially for the complex, uncertain and co-relation elements. Here I abstracted the network into a series of concepts with complex relationships, and scored the concepts and relationships with the help of experts, defined the primary model of the supply chain.Secondly, I take the method of partical swarm optimization. This is one of the random glocal optimization techology, through the interaction of the particals to find the best field of the space, its advantage is easy to realize and have good performance. According to the definition of cognitive map, I prepared 5 groups to be in the steady status and 5 groups to be in the divergent status,totally 10 groups to train the cognitive map, the result map is the pre-alert map that we need.At last, we can base on the optimized supply chain network, input the situation of the concepts, pre-alert the networks with the cognitive map method. The result will tell us the current situation will run into a steady status or a divergent status.