Research On Method Of Case Representation And Retrieval In Emergency Aid Decision-Making System

In recent years, incidents have occurred frequently, which causes governments' great attention. How to make emergency command decision quickly and efficiently after the occurrence of emergency in order to minimize the losses from disasters is widespread concern.Case-based reasoning is a method of analogical reasoning in the artificial intelligence field. By searching similar case to the current issue from case base, choosing one or more high similarity case and modify or adjust it appropriately, the results for solving current issues are obtained. This paper applies case-based reasoning (CBR) approach to emergency aid decision-making, which provides a scientific and effective aid decision-making method for the emergency officials. The article analyzes the current main method of case representation. According to case characteristics in the emergency field, level division method of emergency cases is put forward. Based on the hierarchical structure of emergency cases, a representation method XML is proposed, which solves the unified representation problems of the heterogeneous emergency cases. In the process of studying case retrieval algorithm, single-stage search strategy and the two- stage search strategies based on cluster analysis are designed according to the number of the cases, which improves the efficiency of case retrieval. A two-layer case retrieval algorithm combining structure similarity and attribute similarity is proposed, which effectively overcomes the shortcomings of traditional Nearest Neighbor Algorithm that fails to calculate the similarities between the cases with the missing values. Finally, the realization method of case base in the emergency aid decision-making system is used to illustrate the practicality of CBR approach in emergency aid decision-making.The method of case representation and retrieval in emergency aid decision-making system based on CBR in this article is able to retrieve similar cases to the current issues accurately from historical case base, so as to effectively support decision-makers to make the current issue's decision based on emergency command and dispatch programs of historical cases. A very strong support is provided for decision-makers.