基于概率分布的流程工厂模型拓扑相似度计算  

The Topology Similarity Calculation of Process Plant Models Based on Probability Distribution

作　　者：覃力[1,2] 唐卫清 李士才[4] Qin Li;Tang Weiqing;Li Shicai(Institute of Computing Technology,Chinese Academy of Sciences,Beijing 100190;University of Chinese Academy of Sciences,Beijing 100049;Luoyang Institute of Information Technology Industries,Luoyang 471000;Beijing Zhongke Fulong Computer Technology Co Ltd,Beijing 100085)

机构地区：[1]中国科学院计算技术研究所,北京100190 [2]中国科学院大学,北京100049 [3]洛阳中科信息产业研究院,洛阳471000 [4]北京中科辅龙计算机技术股份有限公司,北京100085

出　　处：《计算机辅助设计与图形学学报》2019年第1期165-175,共11页Journal of Computer-Aided Design & Computer Graphics

摘　　要：为了解决流程工厂模型拓扑相似度计算问题,提出基于概率分布函数的相似度计算方法.首先利用隐马尔可夫建模及状态序列推断,从工厂拓扑结构中挖掘隐含的拓扑语义;然后通过隐含狄利克雷分配模型将拓扑语义映射成长度固定的特征向量;最后用概率模型描述全体特征向量分布,并用分布函数计算2个拓扑结构间的相似度.依据流程工厂专家设计建模的专业知识及从业经验,验证了该方法的准确性和有效性.To solve the problem of topology similarity calculation of process plant models,a similarity calculation method based on cumulative distribution function has been presented.Firstly,the method mines the underlying topology semantic from topology structure through HMM modeling and state sequences reference; then with the LDA model,the topology semantic is mapped to length fixed feature vector; finally,all the feature vectors are described by a probabilistic model and the similarity between two topology structures is calculated based on the cumulative distribution function.According to the specialized knowledge and experiences from process plant professionals of model designing and modeling,the accuracy and effectiveness of the presented method is verified.

关 键 词：流程工厂 拓扑结构 概率分布 相似度计算 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]