基于概率时间Petri网的流程推荐方法

Process recommendation based on probabilistic time Petri net

为了向用户推荐结构相似且时间效率较高的流程,提出了一种基于流程中活动发生的概率和时间的流程推荐方法。定义了一个模型PTN(probabilistic time Petri net)来表示流程,改进了一个现有的流程相似度方法并且将它命名为MDS(matrix distance similarity),用于找出流程库中与查询模型结构相似的流程集合,提出了一个最小加权时间方法(minimum weighted time,MWT),用于找出流程集合中时间效率较高的业务流程。在此基础上设计了相关算法,并且分析了时间复杂度。实验数据证明了方法的有效性和高效性。

To recommend process that is structural similar and has a higher time efficiency to user,this paper proposed a new recommendation method by considering occurrence time and frequency of every activity.First,it defined a probabilistic time Petri net(PTN)to express the time and probability constraint.Secondly,it improved a current process similarity method and further named by matrix distance similarity(MDS)to find the structurally similar process set(PS)to the given query process from the process repository.Thirdly,it proposed a minimum weighted time(MWT)method to find the most efficient process from the PS.Experiment proves the methods’validity and high efficiency.