| In recent years, the railway emergencies take place now and then, which have caused enormous loss to railway transportation. With the continuous application of new technologies and equipment in the field of railway, the complexity and correlation between systems has been deepen, which has intensified the urgency and complexity of response to the emergencies. The traditional emergency management based on emergency plan management is now facing the unprecedented challenges in the field of railway emergencies. The railway emergency management is transforming from the emergency plan management to the scene management. The decision-making model of railway emergency based on scenario and response has been deeply researched to establish the architecture of scenario description of the railway emergencies, and the theories and methods of the scenario reconstruction model of the railway emergencies have also been researched in the dissertation, which has a very important theoretical and practical significance.After the summary of the relevant research results at home and abroad, starting from the deep analysis of the structure of railway emergency scenario, many research methods, involving information science, ontology, evolution analysis, the combination of quantitative and qualitative analysis, the scenario construction of railway emergency and decision has been established. The construction of the scenario model, the integration of scenario information, the emergency and decision, and scenario deduction have been deeply researched in the dissertation.Firstly, the characteristics and process of railway emergency management have been analyzed in detail under the architecture of railway emergency management. The factors have been classified. Thus, the hierarchical network model of railway emergency scenario has been set up through the ontology design of the scenario essentials. The ideas of scenario reconstruction have been proposed to research the process and methods of the scenario reconstruction of railway emergency. The expression of the dependence relationship and visualization of the nodes among the scenario network has been described and analyzed.Secondly, the sources and characteristics of information demand and information scenario construction are analyzed; the railroad emergency information service integration method is put forward based on SOAACM, for the autonomy requirements of railway emergency environment on the information service integration. Through the introduction of Agent technology, the software Agent proxy Web service, the division of scene model globally defined as local process model and various departments makes the information integrated service control to be distributed to multiple service nodes. Compared with the BPEL4 WS structure, information service integration system is more autonomy and scalability by the method.Then, the emergency decision making method based on scenario analysis of railway network is put forward, the scene model, the rough set theory, the D-S evidence theory and Bayesian network are combined together to complete the railway emergency decision reasoning. In this dissertation, from the view of practical application, Bayesian network model is constructed by expert knowledge fusion based on the relation between the factors. In order to reduce the complexity of input parameters scenario network, the information entropy of rough set attribute reduction method is proposed to complete the reduction of the scenarios information set. The results show that the Bayesian network based on the simplified model of decision reasoning algorithm, convenient, can obtain satisfactory results even under imperfect information, more suitable for practical application.Furthermore, the railway emergency scenario deduction model is constructed based on event driven. On the basis of the process of railway emergency scenario evolution, divide the emergency scenario into the initial, middle and end scene, and their mutual relations are analyzed. Combined with the technology of Bayesian network inference, a detailed description of the deduction process is realized. For train derailment accidentals as an example, the node variable state is inferred on the scene, results and reality is consistent to prove the feasibility and effectiveness of the method.Finally, the railway emergency scenario reconstruction and aided decision prototype system of RESRADPS is designed, the system’s physical and logical structure and function module structure design, specific implementation of the prototype system based on hardware and software platform and based on analysis of the system function interface and operation results, verify the theory, the method proposed in the dissertation and the construction of system correctness, feasibility and validity. |
