【摘要】 针对受多种不确定性因素综合作用的流域防洪工程体系风险评价问题,提出用防洪工程体系风险评价指标样本与评价标准等级之间的多元联系数,定量描述样本与标准等级之间隶属关系的层次性和模糊性,用三角模糊数和多元联系数理论相结合的方法处理风险评价中以区间数形式表示的评价等级标准的模糊性,用三角模糊数定量表示联系数的差异度系数的连续变化过程和各指标权重的不确定性,最后通过随机模拟三角模糊数的方法建立了三角模糊数随机模拟的防洪工程联系数风险评价模型(CN-SSTFN)。研究结果表明:CN-SSTFN模型综合利用了流域防洪工程体系风险评价过程的各种主客观不确定性评价信息,在评价结果可靠性方面,以置信区间形式表示的评价结果比现有方法提供了更多的重要信息,能反映受多种不确定性因素综合影响的流域防洪工程体系风险评价的客观实际情况;本模型的计算方法简便、通用,在各种系统风险等级评价问题中具有推广应用价值。更多还原

【Abstract】 In order to deal with the risk evaluation problem of watershed flood control engineering system affected by many uncertainty factors,multivariate connection number expression of index samples and evaluation grade criterions of flood control engineering system was presented to describe hierarchy property and fuzziness of membership relation between index samples and evaluation grade criteria.A method of combining triangular fuzzy numbers with multivariate connection number theory was suggested to deal with the fuzzy evaluation grade criteria problem denoted by interval number in risk evaluation problem of watershed flood control engineering system.The discrepancy degree coefficients of connection number and evaluation index weights were expressed using triangular fuzzy numbers,then a connection number based risk evaluation method for watershed flood control engineering system using stochastic simulation and triangular fuzzy numbers,namely CN-SSTFN for short,was established using Monte-Carlo method.The research results show that various information of subjective and objective uncertainties in the risk evaluation process of watershed flood control engineering system can be comprehensively utilized using CN-SSTFN,its comprehensive evaluation results expressed by confidence intervals can provide more reliability information than that of the common evaluation methods,which can reflect practical conditions of the risk evaluation problem of watershed flood control engineering system affected by many uncertainty factors.The CN-SSTFN can be applied to comprehensive risk grade evaluation problems of different systems.更多还原