基于同步管理器的过程挖掘算法

Process mining algorithm based on synchronized-manager

强循环、弱循环、跳、不可见任务等复杂结构是业务流程中常见的流程结构,从事件日志中挖掘这些复杂结构是亟待解决的问题。分析了强循环、弱循环、跳、不可见任务的特征并给出了它们的判定定理,在此基础上结合同步管理器模型提出了γ+算法,该算法克服了目前过程挖掘算法的限制,在挖掘流程结构的同时挖掘管理行为,能够处理强循环、弱循环、跳、不可见任务等复杂结构,实例分析表明了算法的有效性与可适用性。

Strong-loop, weak-loop, jump and invisible tasks are ubiquitous between tasks in the business process, and mining them from the event logs is a problem to be solved. The features of strong-loop, weak-loop, jump and invisible tasks were analyzed, and the determinant theorems were offered, on the basis of which the γ^＋ algorithm was proposed with synchronized-manager models. The γ^＋ algorithm overcomes the limitation of the converutional mining algorithms, executes the mining of management actions along with the mining of process structure, and manages to solve problems such as strong-loop, weak-loop, jump and invisible tasks. Experimental results indicate the effectiveness and applicability.