区域洪涝灾害风险的模糊综合评价与预测

以广西1992~2006年洪涝灾情数据为基础,选取适当的评价指标,用基于遗传算法、特征值法和优化算法的投影寻踪法和层次分析法确定各评价指标的分类权重和排序权重,用优化算法进行综合得到组合权重,经过加权求和得出洪涝灾害的模糊综合评价,进而对15个评价对象进行了分类排序,其分类结果很好地反映了洪涝灾情的实际情况。在以上模糊综合评价结果的基础上,利用距离贴近度进一步计算和分析各个评价指标之间的贴近度,并由择近原则确定各个评价指标之间在洪涝灾害评价中的相关程度。同时,建立了基于改进的灰色GM(1,1)的时间响应式的灾变年份预测模型。

Fuzzy Valuation and Forecast of Regional Flood and Waterlogging Risk

Based on the flood data of Guangxi from 1992 to 2006,the appropriate valuation indexes are used to determine the weights for classification and sequences by using inheritance algorithm,eigenvalue method and projection pursuit method and analytic hierarchy process of optimization algorithm.Combined weights are obtained by optimization algorithm.The fuzzy comprehensive valuation of flood and waterlogging disaster is given by weighted sum.The results of classification and sorting for 15 evaluated objects fairly indicate the actual situation of flood.Based on the above fuzzy comprehensive valuation,the closeness degree of valuation indexes are calculated and analyzed by using closeness method.The correlative degree of valuation indexes in flood risk valuation by closeness optimization.At the same time,disastrous year forecast model is worked out based on time response of an improved grey GM(1,1) model.