项目名称: 支持应急决策的气象灾害大数据融合的方法研究
项目编号: No.91546117
项目类型: 重大研究计划
立项/批准年度: 2016
项目学科: 宏观管理与政策
项目作者: 吴先华
作者单位: 上海海事大学
项目金额: 43万元
中文摘要: 日益频发的气象灾害已成为我国可持续发展中急需应对的严峻挑战,目前的气象灾害大数据尚未能有效融合,成为灾害应急决策的障碍。本研究拟利用多学科的累积性知识,创新气象灾害大数据融合方法。特色工作主要有:以气象大数据(硬数据)、灾情数据(硬数据+软数据)、社会统计数据(软数据)为对象,采用本体和映射技术,构建基于本体的气象灾害数据标准体系;研究多源数据的分解与融合技术,构建统一的空间元数据标准和地理参考系统;提出动态寻优子空间方法,实现气象灾害大数据的维数约简;采用深度学习技术,实现气象灾害的分类识别;采取案例推理、规则推理、webGIS和前景理论等,生成、评估和反馈应急决策方案,构建开环的“方案(集)—评估—决策”的生成机制,最后通过构建城市暴雨灾害应急决策平台对以上研究给予实证。本研究能对其他灾害类大数据的融合提供借鉴,提出的应急方案生成机制对于以大数据驱动的管理决策范式改变具有典型示范意义。
中文关键词: 气象灾害;大数据融合;应急决策
英文摘要: The increasingly frequent meteorological disasters have become a serious challenge for China’s sustainable development. But currently they are not effectively connected with the big data of meteorological disasters, which has been the obstacle of disaster emergency decision. This research intends to adopt the cumulative multidisciplinary knowledge to innovate the fusion method based on the big data of meteorological disasters. The featured works include: taking meteorological big data (hard data), disaster data (hard data + soft data) and social statistic data (soft data) as the research object; analyzing the knowledge management and data standard of big data of meteorological disasters from ontology perspective; studying the decomposition and fusion techniques of multi-source big data to build a unified spatial metadata standard and geographic reference system; proposing the subspace method of dynamic optimization to realize the dimension reduction of big data of meteorological disasters; adopting the technique of deep learning to realize the classification and identification of meteorological disaster; adopting Case-Based Reasoning, Rule-Based Reasoning, Web Geographic Information System and prospect theory to form, assess and amend the emergency decision scheme, and to build a open-loop generative mechanism of “Solution Set definitions--Assessment--Decision making” and finally carry out the empirical analysis by taking urban rainstorm disaster as examples. This research provides a reference for the similar researches on big data of disasters and proposes a generative mechanism of emergency scheme which has demonstrative value for the change of management decision scheme driven by big data.
英文关键词: Meteorological Disaster;Big Data fusion;Emergency Decision