Study On Method For Generating Emergency Alternative Of Urban Flood Disaster Based On CBR

Because of the abnormal climate, destructed eco-environment and human activity, flood disaster has become a major threat to the human beings. With the increasing expansion of urbanization, urban flood disaster has changed its mode in our country. Moreover, in the facing of rainstorm and flood, modern cities have become more fragile, and waterlogging by rainstorm has affected transport, public underground buildings and houses in central city seriously. So the efficient way to reduce the cost of incidences is to carry on and implement plans without delay. It's the key issue to do a research on how to make emergency plans in the beginning of urban flood disaster, which has very strong practical guidance to emergency decision of urban flood disaster in the future.Lacking of complete systematic emergency knowledge, short-time decision, decision information defect are difficulties that flood disaster decision-makers always confront. Emergency plan has always regarded as the powerful solutions to solve urban flood disaster. However, emergency plan has problems as difficulty in accessing methods, poor practical operation, overdue modifying period and so on. This thesis has employed Case-Based Reasoning technology into urban flood disaster emergency decision field, and used past emergency cases plans to rectify and substitute emergency plan to generate current emergency alternative quickly. Not only CBR can realize the embodiment of the rescue measures, solve the defect part of emergency decision information, but also has the advantage of incremental learning, which means the CBR system can update new cases constantly, delete useless ones with increasing cases in cases base and it's more quick and efficient to update and improve.With the mainline of four stages of CBR, this thesis has analyzed the forming principle of urban flood disaster, generalized the influencing factors of urban flood disaster, selected framework methods to represent problem domain, solutions domain and emergency effectiveness domain, and establish urban flood disaster system of case attributes. Next according to the problem of attribute value loss, uncertain attribute value and complex relationship between attributes influencing the case similarity calculation, this thesis has proposed a new comprehensive indexing technology based on inductive indexing method and RIMER method, and explained its application in the urban flood solving process. Furthermore, in the cases index process, it has also taken emergency alternative' effectiveness in to consideration, used evidence reasoning method twice to synthetize for evaluation and make index results more rationality and accuracy. And then, research the implementation of cases rectifying and learning in CBR. At last, this thesis has provided a framework of method for generating emergency alternative of urban flood disaster based on CBR.