Multi-Clusters Theory In Flexible Framework For Large Information Systems

With the development of distributed network technologies, computers have become prevalent in almost all society fields. Considerable information network infrastructures for routine businesses have been built up. Advanced information technologies play an important role now in industry productions and markets operations. Based on the information infrastructures, growing economics and technologies ask for more advanced services to achieve competent dynamic gingers. Advanced Just-In-Time constructed business represents one of the hottest topics of present information system research. Due to different aims and fields' specialties, many information systems have come out with various layouts, which are, in terms of technology standards and function frameworks, different from each other. A frequent case go like this: many enterprise built up self application architectures, even each department in a given firm developed its own information system to meet its special requirement. As a result, information applications find their difficulty while trying to implement connection and interoperation. What's even worse is that large fundamental data must be maintained overlapped, which sometimes leads to inconsistency. Share and reuse of management information by other system make sense. To provide real-time advanced services from in being information infrastructures by integrating most legacy systems, intelligent business providing (IBP) desires for new state of art in theoretical information systems framework.Based on the rich fruits of AI research, this paper researches Multi-Cluster theory to reach a flexible framework in large-scale integration for information systems reuse, which is then tested in Railway Intelligent Transportation System, an advanced information and control system in Chinese railway. Multi-Cluster theory reconsiders the structures and models of Multi-Agent System (MAS) from the novel view of social economic organizations, and studies the intelligent negotiation methods on sharing operating resources. Enterprise information systems now meet business management with field knowledge in limited area. With optimized organization for services combination, new businesses can be raised among these functionalities. Multi-cluster system is one or more MASs, with more practical social organization structures than IA's BDI structure that mimics human's spirits world. Taking advantage of business abilities, Multi-cluster system introduces intelligent models of negotiation, coordination and collaboration to exploit such kinds of abilities that it can achieve the expected goal by exempting from mass & redundant investments in new business developments. Following the Contract Net Framework, information clusters exchange business function knowledge as well as resource utilization information. 'Marketer Guard' of the cluster is an important part of knowledge integration and credence mechanisms. So advanced new businesses are remodeled dynamically within complex information environment to get an instant decomposition and reintegration over cluster operation platform. Logic business model can fit into the real physical working layer dynamically and intelligently in a robust mode. Safe, reliable, compatible and open business construction dispatches loose coupled clusters' applications resources according to market behavior, of which the bids actions coordinate the business performance.Multi-cluster theory is mainly applied to the analysis and construction of large distributedcomplex systems. Based on the enterprises' specific aims and relationships, Multi-cluster theory tries to build up the 'real-time firms' or 'intelligent markets' through dynamic business negotiations, so to optimize the functional mapping from business logic to physical infrastructure. In words, Multi-cluster system expands the classic integration theories' abstraction granularities in information system developments, which is flexible in reusing legacy systems and carrying out new information businesses.The multi-Cluster organization comes out with characters of recursion and flexibility. As an abstraction method, it covers not only applied information systems and business construction, but also other functional fields.