Scenario Hierarchical Model Of Unconventional Emergency Based On Granular Computing

Compared with the traditional emergency, not only does the unconventional emergency break out without any omen, but also the evolution of unconventional emergency has no rules to follow. Therefore, predicting the outburst and evolution of unconventional emergency by historical data is unrealistic. So the emergency management for unconventional emergency is changing from " prediction-response " to " scenario-response".The methods for representation of scenario have been given in many papers. The trend analysis of scenario is realized based on existing models of representation. All in all, the previous studies, which provide rich theories, methods and technologies, can give us some enlightenment in representing, structuring and analyzing the scenario of unconventional emergency. However, emergency-decision-making is a process to match the knowledge of with the information that decision-makers have been obtained. Due to the differences in role and environment between different decision-makers, the knowledge they gained is somewhat different in size. So the decision-makers at different levels should be provided different size information. That is to say, the decision-makers in high-level need information that can describe the overall situation of events in macroeconomic, but in lower levels,decision-makers pay attention to specific information on site. Therefore, the scenario information should be distinguished in granularity, the corresponding representation model of scenario should also have the ability to provide different granularity of scenario information for dicision-makers in different levels. Nevertheless, existing studies have not explored from the perspective to accomplish represation and construction of scenario for unconventional emergency.To cover the deficiency of the previous studies, based on the theory of granular computing, many studies have been carried out in this paper. Firstly, construction of scenario hierarchical model. The formal representation of the scenario elements is given at first. Then the granuler layer of scenario is constructed based on the depiction of scenario granulation and the relationship between scenario granulations. Meanwhile, the definition of granular degree is given. Further more, this paper builds the mapping between granuler layer on the basis of partial ordering relation. Based on the above description, scenario hierarchical model is established. Secondly, scenario granulation computing for multi-level users. Depend on the mapping between the granular layer, the algorithm, which can provide scenario information for multi-level decision-makers in different granularity, for generalization and refinement of granular layer are carried out. Besides, To support the generalization process between granular layer, intergranular reasoning algorithm in one granular layer is given.Thirdly, using the debris flows as a example, combined with the prototype system, the process of how to get multi-level scenarios and how to provide multi-granularity scenario information on the basis of multi-level scenarios is elaborated clearly.The study in this paper is significance for unconventional emergency management. Firstly, combined definition of scenario in unconventional emergency with the theory of granulation computing, scenario granulation is constructed to solve the problem of how to represent scenario information. Secondly, representation model of multi-level scenario, which provides support for multi-level decision-makers, is given based on the structure of scenario granulation. Thirdly, multi-granularity scenario information can be acquire through the algorithm of generalization and refinement. In the end, all the theories and methods proposed in the paper are implemented through a prototype system.