应急预案计算机辅助生成若干关键问题研究

Emergency plan construction is the core of improving the ability of emergency management of natural disasters. So far, the researchers interested in disaster emergency plans have given a lot of attention to the basic structure of plans and plan evaluation, there is less work on big data knowledge-based reasoning to construct emergency plans aided by computers. At first,in this project we explore the formal model of emergency plans to propose a fine storage mode of emergency plans.On this basis, we can establish the flexible selection framework and processes for emergency plans, integrated into the sifting guidance of experts, combined with the idea of collaborative filtering algorithms, to achieve accurate sifting of cases. Then, the adaption and adjustment of emergency plans through rule-based reasoning, and the expansion of the case and rule sets are researched, the rule management and maintenance mechanisms and the rule set/case set refining algorithm are proposed. Next, a digital plan produce system aided by computers integrating big data knowledge-based reasoning is built, to promote the construction of scientific, operable, flexible and easy to use emergency plans.

应急预案建设是提高自然灾害应急管理能力的核心。现有自然灾害应急预案方面的研究更多关注预案制定原则、基本结构及预案评价，利用大数据知识推理实现应急预案计算机生成方面的研究尚未见到。本项目首先探索自然灾害应急预案形式化模型，提出自然灾害应急预案精细化表示模式。在此基础上建立自然灾害应急预案要素的柔性筛选框架和流程，融入专家筛选指导，结合协同过滤思想，实现精准筛选。进而，研究基于规则推理的案例调整与改写、规则库/案例库膨胀，提出规则库/案例库精炼算法。然后，集成这些研究成果，利用大数据知识推理构建应急预案计算机生成系统，促进科学、可操作、灵活易用应急预案的建设。

应急预案建设是提高自然灾害应急管理能力的核心。现有自然灾害应急预案方面的研究更多关注预案制定原则、基本结构及预案评价，利用大数据知识推理实现应急预案计算机生成方面的研究尚未见到。本项目首先探索了自然灾害应急预案形式化模型，提出自然灾害应急预案精细化表示三层模式。建立了自然灾害应急预案要素的柔性筛选框架和流程，融入专家知识，结合协同过滤思想，可实现较精准筛选与推荐。针对应急知识收集和管理，从知识众包和信息抽取两个方面展开研究，初步构建了知识众包与管理平台，提出了众包成本优化的遗传算法，以及面向众包质量控制的人员质量评价方法和激励机制。研究了基于规则推理的案例调整与改写、规则库/案例库膨胀，提出了规则库/案例库精炼算法。集成这些研究成果，利用大数据知识推理构建了应急预案计算机生成原型系统，可促进科学化、可操作、灵活易用应急预案的建设。