摘要
针对现有过程模型行为相似性度量算法计算结果不够精准或者计算效率较低等问题,提出一种基于可覆盖图编辑距离的行为相似性度量算法(TAGER)。TAGER算法通过定义变迁标签图来表达过程模型的行为,定义了考虑模型行为的图编辑操作以及编辑距离的计算方法,设计了A*算法计算变迁标签图节点间匹配的最优方案。通过三角不等式满足率和平均计算耗时对TAGER算法的表现进行了评估,并从计算效率及性质满足情况等方面与其他主流算法进行了对比。
Aiming at the problem that the calculating result of behavioral similarity measures was inaccurate and inefficiency in existing process models,an approach named Transition-labeled Graph Edit Distance（TAGER）to compute the similarity based on edit distance between coverability graphs was proposed.Through defining the transition-labeled graph,the behaviors of process models was expressed.The computing method of graph edit operations and graph edit distance according to behavioral features was refined.The matching optimal scheme between nodes of two T-labeled graphs was calculated with A＊algorithm.Through the satisfied rate and average time cost of triangle inequality,TAGER from efficiency and efficacy perspectives was evaluated,and an experimental comparison between TAGER and other mainstream similarity algorithms was completed.
出处
《计算机集成制造系统》
EI
CSCD
北大核心
2016年第2期343-352,共10页
Computer Integrated Manufacturing Systems
基金
国家自然科学基金资助项目(61472207)
国家自然科学基金杰出青年基金资助项目(61325008)~~
