【摘要】 基于大坝安全监测数据及专家评价等重要信息源，构建了改进型D-S证据理论。通过加权优化的D-S证据理论处理监测信息，采用群决策法确定专家信息源的基本概率赋值，扩大融合的底层指标数来降低融合结果的模糊性，再采用VIKOR法修正信息源的基本概率赋值，最终运用D-S证据理论融合得到评价指标的安全性态值。结果表明，多源信息证据融合结果相较于单一信息源证据结果的不确定因素系数更小，用VIKOR法给两个证据源择优出关键证据并赋予权重，使得融合结果具有更高的可信度。更多还原

【Abstract】 Based on the important information sources such as dam safety monitoring data and expert evaluation, an improved D-S evidence fusion theory is constructed. Through processing the monitoring information by weighted and optimized D-S evidence theory, identifying the basic probability assignment of expert information source by group decision method, reducing the fuzziness of fusion results by expanded and integrated underlying index number, and correcting the basic probability assignment of information source by VIKOR method, the safety state value of evaluation index is finally obtained by using the D-S evidence theory fusion. The results show that the uncertainty factor coefficient of the multi-source information evidence fusion results is smaller than that of the single information source evidence, and using VIKOR method to select the key evidences from the two evidence sources and assign weights to them makes the fusion results more reliable.更多还原