异构图形数据的近似语义匹配技术研究

以流程工厂协同设计应用为背景，提出基于允许误差的最大语义图匹配（MSMGE）算法的异构图形数据近似语义匹配模型。利用类无向图来描述2D和3D异构图形数据的工程属性和拓扑关系，消除了图形信息的异构性，并建立各种类实体的属性标签词典来消除2D和3D属性信息的异构性，用语义表达式来表示类无向图顶点和边的语义关系，将异构图形匹配转化为近似语义图匹配。通过基于工程语义对类无向图进行语义分割和基于最大公共序列算法的语义表达式比较、语义规整和语义裁剪等方法，降低了匹配搜索空间，提高了近似语义图匹配效率，实现了近似语义图匹配判断。该研究已经在流程工厂设计软件中得到较好地应用。

Algorithm for semantic approximately matching of heterogeneous graphic data

Based on the computer ing of heterogeneous graphic data presented.The undirected-graph-like 3D graphic data to smooth away supported cooperative in process plant design,the model of approximately semantic match- based on Maximum Semantic Matching of Graphic with Error （MSMGE） algorithm is is proposed to illustrate the engineering properties and topological relations of 2D and heterogeneous of graphic data.And then the lexicon of entities is constructed to smooth away heterogeneous of properties,and semantic expression is adopted to represent the semantic relations among vertexes and edges in the undirected-graph-like.So the matching of heterogeneous graphic data is transformed into approximately semantic graphic matching.To improve the efficiency of approximately semantic graphic matching algorithm,reduce the searching space, and realize the approximately semantic graphic matching, some algorithms are adopted, such as the algorithm of graphic parti- tioning, the algorithm of semantic representation matching based on largest common subsequence algorithm, semantic normal and semantic prune,etc.The research has been well applied in the process plant design soft.