Process models constitute valuable artifacts for organizations. A process model formally captures the way an organization works internally and interacts with its customers and partners. Over time, more models may be created as business practices evolve (leading to different versions of models) or an organization expands, e.g., through mergers or acquisitions. It is not uncommon for large organizations to have to manage thousands of process models. Retrieval of process models with desired properties then poses a significant challenge, particularly when one is concerned with finding models that describe certain process scenarios, i.e., sequences of tasks captured in models. This paper proposes a method for automated process model retrieval based on scenario compatibility. A process model is retrieved if it has the potential to perform the specified process scenario. To allow for scenarios to be underspecified, wildcards may be used in their description. The paper reports on a formal language for scenario-based process querying, its implementation, and evaluation in the context of industrial and synthetic process models. The results show that the technique works in (close to) real time.

流程模型构成了组织的宝贵工件。流程模型可以正式捕获组织内部工作以及与客户和合作伙伴进行交互的方式。随着时间的流逝，随着业务实践的发展（导致不同版本的模型）或组织的扩张（例如通过合并或收购），可能会创建更多模型。大型组织必须管理数千种流程模型并不少见。因此，具有所需属性的流程模型的检索带来了巨大的挑战，尤其是当人们关注寻找描述某些流程场景的模型时，即模型中捕获的任务序列。本文提出了一种基于场景兼容性的自动化过程模型检索方法。如果流程模型有可能执行指定的流程方案，则将检索该流程模型。为了允许场景不明确，可以在描述中使用通配符。该论文报告了一种正式的语言，用于基于场景的过程查询，其实现以及工业和综合过程模型中的评估。结果表明，该技术可以实时（接近）地工作。