Hotel Customer Relationship Management System Based On Multi-agent Coordination Mechanism

CRM(Customer Relationship Management) is that enterprise aim at winning customers, keeping customers, elevating customer' s fidelity and improving profits base on the process of managing interaction activity between enterprise and customer, CRM is a set of advanced ideas, methods and solutions, which provide a kind of capability of creating good product and good service, reacting in time and elevating customer' s satisfaction. CRM can help to find out and lock the best customer, and satisfy their needs and fulfil their wish effectively on the right prices and the right time, throughing the right way and providing the good product and service, because of lacking capacity of discover useful information which conceal in datas, enterprises have no way to convert datas into knowledge, it is impossible to help enterprises to improve their profit by using the valid, unknown and understanding information extracting from huge data base, but, in recently, with the increasing development of data mining technology, everything become into reality, also CRM become one of the most developing area in the data mining application.Though all sorts of data mining system has its own advantage now, it is still not so much intelligence and needs to be improved, agent has greatly intelligence in the artificial intelligence area, intelligentize of data mining process can be realize by combining agent with data mining technology.JunShan hotel is a distributed hotel system in geography, this paper put forward a hotel CRM distributed data mining system model base on multi-agent system, taken every subhotel in this system as a agent, naturely, this model introduce agent' s feature to the process of data mining, makeing data mining system has its own autonomy, reactivity, interactivity and adaptability etc, and help to combine the local knowledge into global knowledge.In data mining aspect, this paper built a hotel CRM oriented data mining model base on JunShan hotel data warehouse combine decision tree with neural net data mining technology, and mined potential coustomer' s featurepattern from the model to forecast customer' s behavior, help hotel decision maker to adjust marketing strategy, reduce risk and make the right decision.In multi-agent platform aspect, This paper realized a ring of star topology fault-tolerance agent platform included three main container nodes base on JADE (Java Agent DEvelopment Framework) software frame, when one of the main container running agents cannot be accessed, other main container in this platform can detect it, and rebuild a new agent platform according to address notification service provided by JADE.