摘要
由于军事系统中存在不同的体制编制和多变的网络结构,并且军事单元能力发挥不稳定,信息延时不确定,因此为完成特定军事任务,系统中会存在诸多OODA过程,以至于难以评估军事系统的效能。针对目前效能评估方法人为定性评估成分较大,或者评估过程忽视军事实体间的相关联系而仅在统计意义上进行评估,将军事系统的不确定性纳入评估体系,基于统计关系学习框架Markov logic以OODA军事策略为基础对不同军事规则进行一阶逻辑建模,并以当前网络态势为先验条件,应用Alchemy工具对表示最终作战效能的原子谓词进行条件概率推理,求得该网络作战效能期望,突出了模型对不确定信息的处理能力。最后研究了军事单元和军事体制对系统效能的影响,实验结果验证了本模型的有效性,表明其具有一定的实践和理论意义。
Since there are different systematic structural formations and changeable network struc- tures in a military system, the performance of military units are unstable and the information delay is un- certain, to fulfill the special military mission, the system will include so many observe,orient,dacide and act(OODA) processes that it is difficult to assess the effectiveness of the military system. In view of the present evaluation methods with more qualitative component, or the ignoring of the relevant contact be- tween military entities of the evaluation process just in statistical sense, the uncertainty of military system is incorporated into the evaluation system. Based on statistical relational learning framework Markov logic different military rules in first-order logic are modeled by taking OODA as the basic military strategy, and with the current network state as prior conditions, Alchemy tools are applied for atomic predicates condi- tional probability reasoning which mean the finally operational effectiveness to get the network operational effectiveness expectations. It is prominent that this model has the ability to deal with uncertain informa- tion. Finally, the impacts of military units and military structure upon the performance of the system are studied, and the experimental results show the validity of the model, indicating that it has some practicaland theoretical significance.
出处
《现代防御技术》
北大核心
2012年第6期87-92,共6页
Modern Defence Technology
基金
国家自然科学基金(71031007
71001105)