考虑公众参与的重大突发环境事件应急决策方法研究

The occurrence of major environmental incidents has the characteristics of randomness, uncertainty and high destructive consequences. To respond to the urgent relief demands, it is meaningful to integrate public participation into the emergency decision-making process, which can help experts to make an effective emergency response plan. However, the evaluation information provided by the public has the characteristics of being large-scale, low reliable and conflicting, thus traditional models cannot be applied to information fusion. Therefore, this project aims to study emergency decision-making problems considering public participation. Firstly, we will improve some clustering algorithms and construct some new reliability measure models for information selection. Then, we will build new conflict degrees and some conflict optimization models based on public satisfaction to deal with public preferences and expert preferences. Thirdly, we will establish a new index of risk factors to measure the risk in the fusion process, and then propose some decision-making methods based on the risk optimization for preferences fusion between the public and experts. Lastly, we will give some applications in the cases of practical major environmental incidents, which will provide a scientific basis for the emergency decision-making considering public participation.

重大突发环境事件的发生具有随机性、不确定性、高破坏性等特征，面对减灾救灾任务的紧迫性，有效利用社会力量参与应急决策，并协助专家制定有效的应急方案，具有重大意义。但是由于公众偏好信息规模大、可信度低且具有一定冲突性等，传统方法并不能有效融合公众的偏好信息、公众与专家的偏好信息。本项目旨在研究公众参与的应急决策问题，首先，针对公众偏好信息的筛选问题，拟改进新的聚类算法，并构建公众偏好的可信度测度模型。其次，针对公众偏好与专家偏好的冲突问题，拟构建新的冲突测度模型，并建立基于公众满意度的冲突优化模型。接着，针对公众偏好与专家偏好的融合问题，拟构建新的风险测度模型，并建立基于风险优化的方案选择方法。最后以实际重大突发环境事件案例为背景进行应用，为公众参与应急决策工作提供科学依据。

重大突发环境事件的发生具有随机性、不确定性、高破坏性等特征，面对减灾救灾任务的紧迫性，有效利用社会力量参与应急决策，并协助专家制定有效的应急方案，具有重大意义。但是由于公众偏好信息规模大、可信度低且具有一定冲突性等，传统方法并不能有效融合公众的偏好信息、公众与专家的偏好信息。本项目对公众参与的应急决策问题进行了研究。首先，针对公众偏好信息的筛选问题，我们研究了一种新的语言偏好信息来刻画公众和应急决策专家表达的不确定性，并将该语言偏好信息转化为云模型进行量化处理，进一步研究了云模型的相关运算法则、得分函数等。针对公众偏好与专家偏好的冲突问题，我们对应急决策专家的非合作行为进行了研究，对多种非合作行为进行识别和协调；同时对应急决策专家的权重研究，构建了一种基于信任关系的多层权重求解模型。针对公众偏好与专家偏好的融合问题，我们对应急决策方案的稳健性进行了探讨，利用参数对排序结果可靠性进行了研究。这些研究能为公众参与应急决策工作提供科学依据。