基于增广Petri网生成受控日志

Generation of controlled logs based on extended Petri Net

为减少冗余日志,降低事件约束不可控对算法评估及验证的影响,提出基于可达状态的随机选择生成受控日志的方法。利用增广Petri网为系统建立模型,依据模型中库所与变迁的结构关系及标识分布构建输入矩阵;基于Petri网可达状态分析方法,随机选择触发可发生变迁,记录变迁序列;对记录进行受控分析,拼装生成多重集事件日志和XES标准日志。实现相关日志生成工具,利用BPIC2020数据进行实验,与已有工具进行比较,实验结果表明了工具的效率和有效性。

To reduce the redundant logs and the impact of uncontrollable event constraints on algorithm evaluation and verification,a method for generating controlled logs based on random selection of the available states was proposed,and the extended Petri Net was used to build a model for the system,and an input matrix was constructed based on the structural relationship between the library and the transformations and the token distribution in the model.Based on the available states analysis method of the Petri Net,random selection triggered the occurrence of transitions and the transition sequence was recorded.Controlled analysis of the records and collocation was carried out to generate multiple set event logs and XES standard logs.The relevant log generation tool was implemented,and experiments were conducted using BPIC2020 data,in which existing tools were adopted for comparison.Experimental results show the efficiency and effectiveness of the proposed approach.