Abstract: In the future complex battlefield environment, objective threat level of different target types are difficult to estimate for the uncertainty of information. Aiming at this problem, an information system of threat assessment is built by the theory of granular computing. Then the knowledge in the form of minimized production rules that contain the expert experience is mined by decision logic language. Furthermore, through the analysis of knowledge uncertainty, a strategy of knowledge reasoning is put forward. In this way, the threat of multi-types targets in the complex environment can be effectively evaluated.

Key words: threat assessment, knowledge obtaining, knowledge reasoning, information system, conflict resolving

摘要： 未来复杂战场环境下信息具有高度不确定性，对于不同类型的目标很难客观地估计其威胁等级。针对该问题，采用粒计算的有关理论建立了可实时更新的威胁估计信息系统，基于决策逻辑语言提取出极小化的规则集，它反应了信息系统所包含的专家经验知识。通过分析知识的不确定性，并给出其不确定性表示，提出了相应的知识推理策略，从而可以对复杂情况下的不同类型多目标进行有效的威胁估计。

关键词: 威胁估计, 知识获取, 知识推理, 信息系统, 冲突消解