基于遗传-层次分析法的堤防安全综合评价  

Comprehensive Assessment of Levee Safety Based on Genetic Algorithm and Analytical Hierarchy Process

作　　者：康业渊[1,2] 苏怀智[1,2] 马文丽[3] 

机构地区：[1]河海大学水文水资源与水利工程科学国家重点实验室,江苏南京210098 [2]河海大学水利水电学院,江苏南京210098 [3]水资源高效利用与工程安全国家工程研究中心,江苏南京210098

出　　处：《人民黄河》2014年第2期9-12,19,共5页Yellow River

基　　金：江苏省杰出青年基金资助项目(BK2012036);水利部公益性行业科研专项(201301061);国家自然科学基金资助项目(51179066)

摘　　要：基于堤防工程安全影响因素多、关系复杂的特点,在对堤防安全综合评价指标体系构建的基础上,应用层次分析法、遗传算法以及集值统计法,构建了堤防安全综合评价模型并探讨了其求解算法。将遗传算法与层次分析法结合起来,研究了判断矩阵一致性检验和堤防安全主要影响因素筛选与权重确定问题;根据专家经验及其对定性指标的评分,运用集值统计法获得其评价值,评价值加权组合所得结果即为堤防的安全评价结果。将该模型应用于某湖泊堤防,得出该堤防的安全评价值为68.69,即该堤防属于轻度危险,从而提出在堤防日常管理中要加强巡查,并在汛前对重点堤段进行维护,发现问题及时解决。According to the characteristics of levee safety of lots of influence factors and complex relationship, comprehensive application of analyti- cal hierarchy process, genetic algorithm, centralization statistical method and so on, the construction theory and implementation algorithm of levee safety assessment model were discussed, based on the construetion of evaluation index system. Consistency test of judgment matrixes, sereening main influence faetom in levee safety and determination their weights are studied. According to the scores of qualitative indices from experienced ex- pelts, the evaluation value could be obtained by centralization statistical method, and then the evaluation result of levee safety was gotten based on weighted eombination. The model was applied to a lake levee, and the safety evaluation value was 68.69, belonging to slight danger. Consequently, it is proposed that strengthening inspections is necessary in daily management of levees, some key parts need maintenance before flood season, and problems finding in flood season should be solved without delay.

关 键 词：堤防工程 安全评价 层次分析法 遗传算法 集值统计法 

分 类 号：TV871.3[水利工程—水利水电工程]