Font Size: a A A

Research On Multi-agent-based Supply Chain Management

Posted on:2002-10-20Degree:MasterType:Thesis
Country:ChinaCandidate:H ZhangFull Text:PDF
GTID:2208360032454348Subject:Computer applications
Abstract/Summary:PDF Full Text Request
Distributed artificial intelligence, an important branch in artificial intelligence field, mainly research on intelligent.. simultaneous.. autonomous and self-autonomous multi-agent system ,it also study Agents how to mutual -. coordinative and collaborative solve problems, In which Multi-agent system Research is hotspot and difficulty. Following network popularization, many researchers more and more arouse the development of Electronic Commerce in twenty-first century. They hope can make greatly development for goods flow and information flow which consist of factory.. supplier.. retailer and users, it is also called supply chain management. My paper just based on above to expand in many fields. Produce order.. receive order. complete order, transfer product.. and service all compose a cooperative relation which is called supply chain. They can connect factory.. supplier.. retailer and user ,integrate their knowledge and resources, form a group so as to develop and allocate product. The base model of supply chain management can be simple divided into "push" and "pull". but this mode is selfish, so it cannot use resource greatest and cannot coordinate benefit well, and maybe waste resource, make the relation of supply unbalanced and reduce the whole benefit. In my paper, author firstly sketch the current research status at homeland and abroad, introduce meaning of paper and some theory support to paper. Following, we explorer the character and performance of single agent and multi agent to establish an third layer model for supply chain management. We divided it into third layer: user layer.. supply chain layer and factory layer, each layer is composed by some agents which have especially function. Then we give elaborate explanation about these agents, especially construct factory agent. We also design negotiation domain servers between user layer and supply chain layer. It can proper process bidding and auction in user and factory group. At same time, there is a typically supply chain relationship in supply chain layer to accomplish some task. The supply chain relations' form, which has many different user and factory, can not only use the tradition mode. if we hope it have better benefit. So in paper, we discuss the strategy that can be used form the supply relation. At first, there is a simultaneous ascending strategy. Under this method many different factory agent can use the bidding policy which allow agent propriety advance price, at same time there is auction mechanism which can guarantee the whole agent system benefit and single agent benefit. But in some circumstances, there also happen abnormity, some agents reach commitment, but there is not any Competitive Equilibrium, so in such status, among agents need decommitment and use other strategy. We also suggest a one-shot strategy for this circumstance. Elaborate discussion on strategy and method can be found in paper. At last, there.are some simple simulative experiments for simultaneous ascending strategy and one-shot strategy, gain some data and result for paper. Then we give out compare.. analyses and discussion, explain the efficiency of two strategy. The conclusion and prospect are in end.
Keywords/Search Tags:Distributed Artificial Intelligence, Multi-Agent System, Supply Chain, Management, Three layer Model about Supply Relation, Negotiation, Task, Deyendency, Network, Competitive Equilibrium, Auction Mechanism, Bidding Policies
PDF Full Text Request
Related items